Role of HBsAg levels in guiding hepatitis B virus prophylaxis in pregnancy: Insights from a multi‐ethnic cohort

Pregnant mothers with chronic hepatitis B infection (CHB) need peri‐partum antiviral prophylaxis (PAP) to reduce the risk of mother‐to‐child‐transmission. Currently, PAP is recommended in those with high viral load (VL) that is, HBV DNA >200,000 IU/mL. Quantitative hepatitis B surface antigen (qHBsAg) >10,000 IU/mL, a cut‐off derived primarily from hepatitis B e‐antigen (HBeAg) positive antenatal cohorts in Chinese populations, is advocated as a surrogate marker of VL for guiding PAP. We investigated the utility of qHBsAg to predict high‐VL in a multi‐ethnic urban cohort with CHB. A consecutive cohort of women with CHB was identified from Barts Health NHS Trust databases in the United Kingdom. We included women with paired HBV DNA and qHBsAg during pregnancy. Women already on antiviral at conception were excluded. A total of 769 pregnancies in 678 CHB pregnant mothers (median age 31 years‐old, 8.6% HBeAg+) were included. At median gestational age of 15.3 weeks, HBV DNA was 336 (IQR 44–2998) IU/mL, with 65 (8.5%) being high‐VL. Serum qHBsAg was most useful in Black/Black‐British/Caribbean/African (AUROC 0.946) with 100% sensitivity and 80.6% specificity to predict high‐VL; but it performed less well for other ethnicities: Asian (AUROC 0.877), White (AUROC 0.797) and mixed ethnicities (AUROC 0.742). In conclusion, for settings where healthcare resources are not limited, HBV DNA remains the optimal marker to identify highly viraemic pregnancies for guiding PAP. For resource‐limited settings where the prevailing cost is treatment, serum qHBsAg can be used in Black/Black British/Caribbean/African sub‐cohorts, but not for other ethnicities.


| INTRODUC TI ON
Chronic hepatitis B (CHB) is a major global health challenge.The World Health Organization (WHO) have estimated that over 290 million people are infected, and this leads to 820,000 deaths per year. 1 In light of this, WHO have set out to eliminate viral hepatitis as a major public health threat by 2030.Hepatitis B virus (HBV) is endemic in large geographic regions with endemicity maintained predominantly through mother-to-child transmission (MTCT), also referred to as vertical transmission.The focus on reduction of new cases must therefore be the reduction of the prevalence of chronic infection in under 5-year-olds. 2st international guidelines endorse that the risk of MTCT is determined by the infectivity of the mother; defining high-risk women as being hepatitis B e-antigen (HBeAg) positive, having HBV DNA >200,000 IU/mL or quantitative hepatitis B surface antigen (qHBsAg) levels >10,000 IU/mL. 3,4In highly viraemic mothers, the risk of MTCT can approach 90% without timely neonatal immunization of the baby.Giving a dose of HBV vaccine and hepatitis B immunoglobulin (HBIG) to the infant within the first 24 h of birth can reduce the risk of transmission to <10%.Nevertheless, immunoprophylaxis alone does not completely eradicate the risk of HBV transmission. 5][8][9][10][11] Historically, the gold standard for risk stratification in guiding PAP has been to use HBV DNA level.4][5] Given the high cost and poor accessibility of HBV DNA testing in many low and middle-income countries, alternative strategies of PAP other than assessing VL (PAP-VL) have been explored, such as offering PAP to HBeAg+ (PAP-HBeAg) or all HBsAg+ pregnant mothers regardless of other viral markers (PAP-universal). 12cently, studies (predominantly from Asia) have evaluated the utility of qHBsAg as a single marker of risk, [13][14][15][16][17] which has been shown to be a good predictor of high VL.Measurement of serum HBV DNA and HBsAg can reflect similar, but not identical, virological events.In particular, quantification of serum HBV DNA alone reflects viral replication activity; whereas serum HBsAg is produced not only from translated messenger RNAs of transcriptionally active cccDNA, but also from integrated HBV DNA sequences which become the predominant source of HBsAg in HBeAg-subjects. 18,19erefore, the usefulness of qHBsAg in the setting of pregnancy remains ill-defined.
Antenatal screening for HBV is offered to all pregnant women in the United Kingdom under the National Health Services (NHS) with an uptake rate of >99%, 20 which captures most cases of HBV in pregnancy and represents a non-biased cohort in the United Kingdom.In this study we investigated the utility of qHBsAg to predict highly viraemic pregnancy in a multi-ethnic urban cohort with CHB in a large NHS teaching hospital in central London, United Kingdom and explored the role of qHBsAg in guiding PAP.In addition, we assessed whether the HBV DNA profile at the index pregnancy was consistent in subsequent pregnancies.

| Patients.
We conducted a cross-sectional study involving a consecutive cohort of women with CHB, defined as serum HBsAg positivity for >6 months.Between December 2012 and February 2022, this consecutive cohort, was identified from Barts Health NHS Trust databases and collected by the HBV service across the Trust.Paired assessment of serum HBV DNA and qHBsAg was performed as confirmatory tests in women who tested HBsAg positive on initial antenatal screening.Patients were excluded if they were already on antiviral therapy (Figure S1).Ethnic groups are categorized according to the agreed list of ethnic groups in England and Wales 21 as follows: Asian/Asian-British (including Indian, Pakistani, Bangladeshi, Chinese or any other Asian background), Black/Black-British/Caribbean/ African, White (including English, Welsh, Scottish, Northern Irish or British, Irish, Gypsy or Irish Traveller, Roma or any other White background), and mixed/others (including White+Black Caribbean, White+Black African, White+Asian, any other mixed or multiple ethnic background or other ethnic groups not specified above).All research was conducted in accordance with both the Declarations of Helsinki and Istanbul.The study protocol was approved by the Barts Health NHS Trust Audit Office.All authors had access to the study data and reviewed and approved the final manuscript.

| Laboratory measurement
Serum HBV DNA was measured using automated real-time PCR assays on the Roche Ampliprep/COBAS Taqman prior to 2018 and the Roche Cobas 6800 from 2018, with lower limit of quantification/detection (LLOQ/LLOD) of 20 IU/mL and 10 IU/mL, respectively.Both qualitative HBeAg and qHBsAg were measured using chemiluminescent microparticle immunoassays (CMIA) on the Abbott Architect i2000, The LLOQ/LLOD of qHBsAg is 0.05 IU/mL.

| Definition
Baseline was defined at the time of blood taking for qHBsAg and HBV DNA during pregnancy.For mothers with more than one pregnancy, index pregnancy was the first presenting pregnancy, while subsequent pregnancy was defined as any pregnancy following the index pregnancy.Highly viraemic was defined as serum HBV DNA above 200,000 (i.e.5.3 log 10 ) IU/mL 3,4 and this was the threshold to commence PAP (PAP-VL) to reduce the risk of MTCT in this setting.High qHBsAg was defined as levels above 10,000 (i.e. 4 log 10 ) IU/mL.3348 IU/mL and vs. 8204 vs. 4401 IU/mL, respectively, p < .001for both comparisons; Figure 1).

| Performance of qHBsAg in determining highly viraemic pregnancy
The AUROC was 0.834, 0.777 and 0.679 for the overall cohort, HBeAg+ and HBeAg− cohorts, respectively.Among HBeAg+ patients, a cut-off level of four logs was 76% sensitive and 69.2% specific to determine highly viraemic pregnancy.Among HBeAgnegative patients, the same cut-off level was only 46% sensitive and 74.3% specific to determine highly viraemic pregnancy.When TA B L E 1 Baseline characteristics of all included subjects.a cut-off level of four logs was 100% sensitive and 80.6% specific to predict highly viraemic pregnancy.In contrast, qHBsAg performed less well for Asian (AUROC 0.877), White (AUROC 0.797) and mixed ethnicities (AUROC 0.742) (Figure 2).

| Performance of qHBsAg and/or qualitative HBeAg in determining highly viraemic pregnancy
The distribution of serum HBV DNA levels with respect to qHB-sAg and stratified by HBeAg status was demonstrated in Figure 3.   2).

| Comparing qHBsAg and HBV DNA between index pregnancy and subsequent pregnancy
A total of 91 patients (HBeAg+ in 7 patients) had more than one pregnancy (2 pregnancies in 86 patients, 3 pregnancies in 4 patients and 4 pregnancies in 1 patient).No patients changed HBeAg status between the index pregnancy and subsequent pregnancy.
Six patients were highly viraemic at the index pregnancy and were treated with NUCs in line with the current guidance.Their HBV DNA profile remained consistent at the subsequent pregnancy.No patients in the low viraemic group at index pregnancy became highly viraemic during a subsequent pregnancy.The overall consistency of HBV DNA profile across pregnancy episodes was 100%.

| DISCUSS ION
To reduce MTCT of HBV, the PAP-VL approach has been endorsed by the WHO and major international guidelines, where high VL was defined as HBV DNA ≥5.3log 10 IU/mL. 5 In settings where antenatal HBV DNA testing is not available or not readily accessible, the PAP-HBeAg approach is recommended as an alternative. 5Similarly, qH-BsAg is recommended by EASL as a surrogate marker in situations where there are resource constraints. 3Serum qHBsAg positively correlated with maternal VL (r = 0.69; p < .001)and significantly associated with risk of MTCT with a comparable AUROC as HBV DNA. 16e cut-off level of >4.1 log 10 IU/mL was 87.5% accurate to predict HBV DNA ≥7.0log 10 IU/mL. 144][15][16][17] The United Kingdom, similar to other Western countries, is considered to have an advanced healthcare system with no major resource constraints.Our study attempted to explore the approach of PAP-qHBsAg-whether qHBsAg would be of value to guide PAP in such healthcare settings.In particular, our study comprised a large multi-ethnic cohort involving more than 700 pregnancies in women with CHB and involved a representative mixture of different ethnic backgrounds.Within this unique population, we found that serum qHBsAg correlated well with serum HBV DNA (r = 0.784, p < .001),but more robustly in HBeAg+ mothers (r = 0.882; p < .001) than HBeAg-mothers (r = 0.279; p < .001).Our findings are consistent with a Canadian multi-ethnic study (although comprising predominantly Asian patients) which found a significant correlation between qHBsAg and HBV DNA levels primarily in HBeAg+ patients. 22en assessing the performance of qHBsAg as a surrogate marker of significant viremia, the AUROC for the overall cohort was 0.834, whereas less robust performance was shown for HBeAg-individuals.A qHBsAg cut-off level of 4 logs was 76% sensitive and 69.2% specific to F I G U R E 3 Dot plot showing the distribution of serum HBV DNA levels with corresponding qHBsAg levels, stratified by HBeAg status, in all pregnancies (N = 769).HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; qHBsAg, quantitative hepatitis B surface antigen.
determine highly viraemic pregnancy but was less accurate in HBeAgmothers.In our multi-ethnic single-centre study, qHBsAg performed better for Black/Black-British/Caribbean/African patients (100% sensitive with 100% NPV).Although HBV genotype was not available for the majority of our patients, it is well-reported that in many areas of Africa, genotype A is predominant and this may be indicative of a genotype A-predominant HBV infection in our Black/Black-British/ Caribbean/African cohort. 23This may partly account for the highest level of qHBsAg seen in these patients, as this qHBsAg-genotype A correlation is well-established. 24It is evident that our results would not support integration of qHBsAg measurement routinely instead of HBV DNA; but this can be considered in specific settings (see below).
The cost of HBV DNA testing inevitably exceeds that of HBeAg or qHBsAg, or even provision of PAP in resource constrained settings.
In our NHS Healthcare Trust, the equivalent estimated costs for qualitative HBeAg, qHBsAg and HBV DNA are USD $4.6, $8.8 and $55.9, respectively.In analogy to this, as per 2016 data, access to HBV DNA measurement remained limited and unaffordable in most resource-limited and high-prevalence regions (US$100-$400 per test). 25Likewise, a cost effectiveness study from 2016 estimated that the use of HBV DNA as compared to qHBsAg costs approximately USD $20,000 more per infection prevented. 22A more recent cost effectiveness study highlights that the PAP-VL approach is only cost-effective in 26% of low-and middle-income countries due to high diagnostic cost. 12Therefore, the PAP-universal approach has been increasingly considered especially if diagnostic tests to identify high-risk mothers of MTCT are not available. 12The PAP-universal approach has been evaluated by Mokaya et  for South Africa and other low/middle income settings through a decision-analytic model simulating 10,000 singleton pregnancies. 26cently, the Chinese Medical Association updated the CHB management guidelines and recommended to expand the treatment criteria, one of which involves patients age >30 with any levels of detectable DNA. 27If this criteria were to be considered in our cohort, 365/678 (53.8%) subjects would be eligible for antiviral therapy, compared to 8.7% (highly viraemic) or 30.8% (high qHBsAg) using the existing guidelines.This is consistent with the estimation that the use of PAP will be increased by five times compared to PAP-VL approach, 12 which leads to additional considerations if the PAP-universal approach is to be adopted.These include the capacity of antenatal care units to deliver the care, patients' willingness to adhere to treatment, accessibility of stable supply of TDF, integration between healthcare disciplines (midwives/obstetricians and hepatologists) and the potential for hepatic flares following treatment cessation post-delivery if baseline VL is not assessed, to name a few.Moreover, any healthcare system including the NHS would require new and more efficient patient pathways to deliver antiviral therapy at this scale.
Considering our findings, however, in healthcare settings where HBV DNA is affordable, it remains the most important test for risk stratification of MTCT.Additionally, qHBsAg does not seem to be consistently reflective of HBV DNA in our mixed population.We believe that HBV DNA measurement should be regarded as a gold standard to decide whether NUCs should be prescribed during pregnancy, that is, PAP-VL approach.Moreover, HBeAg measurement is a low cost test and forms part of the standard of care for all CHB patients, therefore it should not be omitted from disease assessment.The sensitivity of a diagnostic test to identify highly viraemic mothers should be prioritized over specificity, PPV and NPV in the context of preventing MTCTthe benefits of preventing vertical transmission outweigh the risks of over-treating a pregnant mother with PAP who indeed has low HBV DNA.It is noteworthy that in our multi-ethnic study, qHBsAg was 100% sensitive in Black/Black-British/Caribbean/African patients, while HBeAg is 100% sensitive for patients with mixed ethnicities.In light of this, in resource-limited settings where qHBsAg or HBeAg is more accessible than HBV DNA, we believe that qHBsAg (i.e.PAP-qHBsAg) or HBeAg (i.e.PAP-HBeAg) could be considered to risk-stratify Black or mixed-ethnic pregnant women, respectively.HBeAg should not be used alone to identify highly viraemic pregnancies in White patients due to low sensitivity (42.9%).However, extrapolation of these data to resource-limited settings should be undertaken with caution, and will require further investigation.Figure 4 illustrates the considerations to be given when deciding which strategy for reducing MTCT, including the limitation in healthcare resources, the prevailing cost (diagnostics vs. antiviral therapy), and the ethnicity of the pregnant mothers.

F I G U R E 4
Considerations to be given when deciding which strategy for reducing MTCT.In resource-limited settings, if the prevailing cost is diagnostic tests (instead of treatment) or these tests are not available, PAP-universal approach should be adopted, which refers to treating all pregnant mothers who are HBsAg+ regardless of other viral markers.In contrast, if the prevailing cost is treatment (instead of diagnostic tests), risk stratification approach with PAP-VL (White and Asian subjects), PAP-HBeAg (mixed ethnicities), or PAP-qHBsAg (Black/ Caribbean/African) should be adopted.HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; NUC, nucleoside analogue; PAP, peripartum antiviral prophylaxis; qHBsAg, quantitative hepatitis B surface antigen; VL, viral load.
In resource-limited settings, if the prevailing cost is diagnostic tests (instead of treatment) or these tests are not available, PAP-universal approach should be adopted.In contrast, if the prevailing cost is treatment (instead of diagnostic tests), risk stratification approach with PAP-VL (White and Asian subjects), PAP-HBeAg (mixed ethnicities) or PAP-qHBsAg (Black/Caribbean/African) should be adopted.
Lastly, we assessed the consistency of serum biomarkers across more than one pregnancy per patient.A total of 91 patients fell within this category during the duration of the study.Our study provided evidence that the HBeAg status remains unchanged across pregnancies.Additionally, at a median interval of more than 2 years, qHBsAg and HBV DNA were largely similar at the index pregnancy and subsequent pregnancy, with the exception of qHBsAg in five HBeAg-cases (Table S1).We showed that no patients in the low viraemic group at index pregnancy became highly viraemic at subsequent pregnancy; and the overall consistency of HBV DNA profile across pregnancy episodes was 100%.It appears that in CHB mothers with multiple pregnancies, a one-off measurement of HBV DNA at the index pregnancy (first antenatal visit) may be sufficient in healthcare settings with limited resources, where repeat HBV DNA measurements at subsequent pregnancies would be unnecessary as HBV DNA levels rarely cross category from low to high viremia.On the other hand, as the current HBV treatment paradigm shifts towards an earlier treatment approach, 28 NUCs are likely to be continued beyond the suggested post-partum period of prophylaxis, especially if future pregnancies are planned.
Although our study has several advantages, there are a number of limitations which warrant consideration.First, this a single-centre study, despite the significant number of patients included from an NHS Trust that serves a population of approximately 2.5 million people.
Second, in contrast to other similar studies, the foetal outcome is uncertain.This is due to the fact that following delivery, administration of HBV vaccination/HBIG and discharge from hospital, the newborns are then followed-up and tested locally in a primary care setting.However, in our Healthcare Trust, all highly viraemic mothers were put on antiviral therapy, together with birth-dose HBV vaccination and HBIG administration to newborns.Lastly, in our study, pregnant mothers on antiviral therapy were excluded.In the United Kingdom, it is standard of care to offer antiviral treatment where indicated, regardless of the pregnancy status.This may represent a selection bias as reflected by a relatively low median HBV DNA, in contrast to 'resource-limited' settings where more heterogenous populations are likely to be found, including mothers with highly viraemic CHB infection.
In conclusion, maternal serum qHBsAg can predict highly viraemic pregnancies in Black/Black British/Caribbean/African women within a mixed population in a large NHS Trust in the United Kingdom, but its utility in other ethnic groups with CHB in pregnancy is less clear.Considering this, HBV DNA should be considered the mainstay of screening pregnancies, whereas qHBsAg or HBeAg (which are relatively inexpensive compared to HBV DNA) may have a role in resource-limited settings where the prevailing cost of diagnostics is lower than therapy, depending on ethnicity of the patient.
Additionally, a one-off HBV DNA determination at index pregnancy may act as a surrogate for subsequent pregnancies without necessity for re-testing during antenatal visits, outside the routine CHB follow-up in resource-limited settings.Further studies are required to validate these results.
and mixed ethnicity.In general, qHBsAg alone offers quite low PPV across most ethnic subgroups (7.4%-17.4%)except Asian F I G U R E 1 Serum qHBsAg levels in pregnancy among different ethnic groups.qHBsAg, quantitative hepatitis B surface antigen.F I G U R E 2 AUROC values for qHBsAg to determine highly viraemic pregnancy for different cohorts.AUROC, area under the receiveroperating characteristic curve; HBeAg, hepatitis B e antigen; qHBsAg, quantitative hepatitis B surface antigen.populations (47.8%).In comparison, HBeAg alone was 100% sensitive and specific for identifying highly viraemic pregnancies from mixed ethnicities, compared to 81.4%, 83.3% and 42.9% sensitivity for Asian/Asian British, Black/Black British/Caribbean/ African and White pregnancies.No obvious improvement in performance metric of each diagnostic test could be observed when either one condition was fulfilled for qHBsAg or HBeAg.If both conditions were fulfilled for qHBsAg and HBeAg, the PPV would improve for Asian/Asian British patients but test sensitivity is compromised (Table

All patients (n = 678) High viral load a (n = 59) Low viral load (n = 619)
Abbreviations: ALT, alanine aminotransferase; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; qHBsAg, quantitative hepatitis B surface antigen.Note: Median and interquartile range shown for continuous variables.a High viral load refers to serum HBV DNA >200,000 IU/mL.categorized into ethnic groups, serum qHBsAg was most useful in Black/Black-British/Caribbean/African patients (AUROC 0.946), and

Table 2
summarizes the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of qHBsAg and/ or HBeAg in determining highly viraemic pregnancy.We used four categories to determine sensitivities: (1) qHBsAg alone, (2) HBeAgThe PPV for qHBsAg is relatively low compared to HBeAg (21.1% vs. 80.0%).In contrast, the NPV for both tests exceed 97%.When both tests are used (i.e.qHBsAg and HBeAg), the PPV would be improved to 89.6%, but 26.1% pregnancies would be unclassified according to qHBsAg and HBeAg, leading to 7.5% highly viraemic pregnancies undetected.When divided by ethnic subgroups, qHBsAg alone was 100% sensitive for Black/Black British/Caribbean/African, but only 74.4%, 71.4% and 57.1% for Asian/Asian British, White Comparison of the sensitivity and specificity of qHBsAg and/or HBeAg in identification of highly viraemic pregnancies.
al. suggesting that it may be cost-effective TA B L E 2 Abbreviations: HBeAg+, hepatitis B e antigen positive; HBV, hepatitis B virus; NPV, negative predictive value; PPV, positive predictive value; qHBsAg, quantitative hepatitis B surface antigen.a 2.4% unclassified ethnicity.b Highly viraemic refers to serum HBV DNA >200,000 IU/mL.c High qHBsAg refers to serum qHBsAg >10,000 IU/mL.d Either one condition fulfilled.e Both conditions fulfilled.