Sexual Dimorphism in Chronic Hepatitis B Virus (HBV) Infection: Evidence to Inform Elimination Efforts

Sexual dimorphism in infectious diseases refers to the different infection susceptibilities and outcomes between males and females, and has been described for many pathogens, including hepatitis B virus (HBV). HBV is a substantial global health problem, with close to 300 million people chronically infected, and accounting for a million deaths each year, with an urgent need for enhanced interventions to support progress towards elimination goals. Sexual dimorphism has a strong influence in HBV infection, with males more likely to be exposed, to develop chronic infection, and to suffer from complications including cirrhosis and hepatocellular carcinoma (HCC) compared to females. Different outcomes are driven by differential immune responses, sexual dimorphism of the liver, and androgen response elements in the HBV genome. The impact of sex may also vary with age, with changes at puberty and influences of menarche, pregnancy and menopause in females. In addition, gender has complex influences on education, beliefs, behaviour and access to / engagement with healthcare services, which may contribute to differences in diagnosis and treatment. Interplay between these complex factors, alongside other attributes of host, virus and the environment, accounts for different outcomes of infection. However, gaps remain in our understanding of sexual dimorphism in HBV, and little effort has previously been made to harness this knowledge for translational gains. In this review, we assimilate human and animal data to consider the mechanism, outcomes and impact of sexual dimorphism, and consider how these insights can be used to inform advances in surveillance, treatment and prevention for HBV infection.


Introduction
Chronic infection with hepatitis B virus (HBV) is estimated to affect 257 million people 1 and accounts for an increasing burden of the 1.34 million yearly deaths due to viral hepatitis 2,3 . United Nations Sustainable Development Goals (SDGs) underpin ambitious targets for the elimination of hepatitis viruses, with the Global Health Sector Strategy setting out aims to reduce new infections by 90% and mortality by 65% by 2030 4 . Improving our understanding of the natural history of chronic HBV infection (CHB) is central to inform progress through enhanced evidence-based treatment and prevention of liver disease 5 .
Sex, defined as the biological characteristics that differ between males and females 6 , accounts for significant immunological differences leading to disparities in outcomes for a variety of infectious diseases 6,7 , termed 'sexual dimorphism'. The effect of host sex on outcomes of infection is complex and multifactorial, influenced by genetics, hormones, and environmental exposures 7 , with trade-offs between protective immune responses (leading to clearance or control of infection) and immunopathology (associated with increased severity and duration of disease). Gender, as a societal and behavioural construct, also plays a role through its influence on perceptions, behaviour, and access to healthcare.
Among chronic viral infections, human immunodeficiency virus (HIV) exemplifies this complex picture, whereby females typically have lower titres of HIV RNA than males 8 , with a 5-fold higher frequency of elite control 9 . However, females also show increased immunopathology, associated with an elevated risk of developing AIDS compared to males with the same HIV RNA levels 6 , and a greater susceptibility to infection, with both a biological and societal basis 10,11 . UNAIDS data for 10-19 year olds in 2019 reported 33,000 adolescent girls becoming HIV infected compared to 4,200 boys 12 . Hepatitis C virus (HCV) has a higher prevalence in males, with females being more likely to clear the virus and also suffering fewer complications 13,14 . Sexual dimorphism in HBV has been less rigorously studied, but was first described by Baruch Blumberg in 1972 15 , nine years after his discovery of 'Australia antigen' (now termed Hepatitis B surface antigen, HBsAg), underpinning subsequent consistent observations of an increased risk of chronic infection and its complications among males compared to females.
HBV is a partially double-stranded DNA virus which archives itself in the nucleus of hepatocytes as a covalently closed circular (ccc)-DNA 'mini-chromosome' accounting for persistent chronic infection, that can lead to inflammatory liver disease, fibrosis, cirrhosis and hepatocellular carcinoma (HCC) 16,17 . Viral factors, the liver micro-environment, and host attributes all contribute to sexual dimorphism in CHB, as previously reviewed 13,18,19 . However, HBV infection has been relatively neglected by research, clinical care, public health interventions, and advocacy 20 , and females are specifically under-represented by translational research 21,22 .
In this review, we discuss sexual dimorphism in CHB, considering the relevance of sex vs. gender, and the specific influence of menarche, pregnancy and menopause in females. We consider how an improved understanding of differential outcomes between males and females may (i) underpin new insights into the pathophysiology of liver disease, (ii) improve patient stratification for surveillance and treatment, (iii) inform new approaches to personalised therapy, and (iv) optimise public health measures and resource allocation. This would support advances towards elimination goals.

Epidemiology of HBV infection according to sex
Blumberg's 1972 metanalysis of HBV prevalence included studies in 23 disparate populations, including cohorts with leprosy, trisomy 21, and renal dialysis patients. HBsAg carriage was more prevalent in males in 22/23 of the populations studied 15 . The male:female ratio ranged from 3.58 to 0.855 and, interestingly, was found to be greatest in the 0-19 age groups in all but one of the studies where age stratification was possible.
More recently, the male:female sex ratio in CHB has been reported as 1.2 in asymptomatic carriers, increasing to 6.3 in chronic liver disease and 9.8 in HCC 23 , but these estimates vary between settings. Selected studies reporting the prevalence of HBsAg in males and females are presented in Table 1, with odds ratio of infection as high as 1.9 in males (95% CI, 1.2-3.2) 24 .
None of the metanalyses we identified (Table 1) could use data from all of the studies identified due to poor and inconsistent reporting of disaggregated sex data. They also suffered from substantial heterogeneity between studies and potential bias due to inclusion of specific high-risk populations. It is notable that the data included in Table 1 do not include any representation of African populations, as a result of the lack of robust metadata available from the African continent. For example, large scale HBV metanalyses of HBV prevalence from Ethiopia and Burkina Faso were unable to determine a sex ratio 25,26 . This highlights a consistent difficulty that currently inhibits robust meta-analysis to more accurately delineate sex differences in HBV prevalence.
A key question to be addressed is whether CHB prevalence is lower in females due to less exposure events, lower susceptibility to acute infection, or enhanced clearance compared to males, or a combination of all of these ( Figure 1, columns A, B and C respectively). In part, this could be assessed by comparing the prevalence of anti-HBc antibodies (indicating exposure, with or without clearance) and HBsAg (indicating current infection)

Amendments from Version 2
Version 3 contains an updated version of Figure 3, and a correction to author affiliations.
Any further responses from the reviewers can be found at the end of the article 95% CI -95% confidence interval. HBsAg -Hepatitis B virus surface antigen * Total number tested is not reported in this study, so denominator calculated from study meta-data, but we cannot exclude the possibility of some population groups being represented more than once.
between males and females. In a cohort of 31,990 blood donors from Crete, HBsAg prevalence was 0.41% in males vs 0.28% in females (OR 1.98, 95% CI 1.2-3.2), and exposure was also higher in males than females, with anti-HBc prevalence of 9.16% vs 5.86% respectively 24 . A large-scale metanalysis in 2019 identified no significant difference in spontaneous HBsAg seroclearance between females and males 32 . However, there was substantial heterogeneity between studies and HBsAg clearance is an uncommon event 33 , so the study may have been underpowered to detect a sex influence.
The epidemiological literature for HBV suffers from over-representation of specific populations that are subject to sex or gender bias. For example, certain defined populations, such as healthcare workers, pregnant women, waste handlers, men who have sex with men (MSM), people who inject drugs, and sex workers are often studied 34,35 . The increased prevalence of HBsAg in males may reflect a complex combination of exposure risk, response to acute infection, and outcomes in chronic infection. More epidemiological data are needed, disaggregated by age, sex and disease characteristics, in order to determine local requirements for provision of public health and clinical services, and to build a clear picture of the global burden of disease.

Sexual dimorphism in liver disease
The liver is a highly sexually differentiated organ, with up to 70% of genes showing differences in expression between male and female mice 19,36 . In humans, females are relatively protected from chronic liver fibrosis, regardless of aetiology, including lower rates of NAFLD/NASH 37 , a lower rate of cirrhosis and liver transplant 38 and a lower risk of hospitalisation and death from cirrhosis 39 compared to males. Pre-menopausal females with CHB are at a lower risk of chronic liver disease compared to males, as exemplified by a cohort of 672 patients in Switzerland, in which liver-related outcomes were significantly less common in females (OR 0.35, 95% CI 0.20-0.60) 40 . Similarly, in a Canadian study of nearly 6000 patients with HBV infection, male sex was an independent predictor of advanced liver fibrosis 41 . There is increasing recognition that diabetes and metabolic disease are associated with worse outcomes in HBV infection 42 , which may be linked to increased disease in males, as differences in fat biology and metabolism differ by sex 43 . Metabolicassociated fatty liver disease is also significantly more common in males, with a loss of protection in post-menopausal females 44 .

Hepatocellular carcinoma (HCC)
HBV is the single biggest aetiological agent of liver cancer worldwide, causing over half of all cases (point estimate 56%, 95% CI: 52-60), but with substantial regional variation, such that in parts of sub-Saharan Africa and Eastern Asia it is responsible for at least 2/3 of cases 45 , and far more in some settings. Overall, the incidence and mortality of liver cancer is higher in men than in women ( Figure 2). A Global Burden of Disease study estimates annual incidence of primary liver cancer at 591,000 in males vs. 264,000 in females (ratio 2.2); among these, HBV-related HCC was estimated to have caused 203,000 cases in men and 70,000 in women (ratio 2.9) 16 . In specific studies, male:female ratios for HCC are between 2:1 and 9:1. This wide variability is at least partly accounted for by the prevalence of CHB 46 , as in settings in which HBV accounts for a high prevalence of HCC, the male:female ratios are typically high 47 , for example in Senegal, where HBV accounted for almost 70% of HCC cases, the sex ratio was 6.6 48 , and in Vietnam, where >80% of HCC was HBV-associated, the ratio was 8.9:1 49 . HCC outcomes are worse when the aetiology is HBV 50 , and in males compared to females 51 .
These observations in humans are corroborated by evidence from animal models that explore sex-related susceptibility to the development and progression of liver cancer, irrespective of aetiology. For example, following exposure to high dose carcinogens (such as diethylnitrosamine (DEN)), liver inflammation and proliferation are more pronounced in male mice 52 , and males exhibit a higher rate of tumorigenesis 54 , with cancers that progress more rapidly.

Flares of liver inflammation and acute-on-chronic liver failure
Flares of liver inflammation in HBV infection can be a consequence of the immune response targeting infected hepatocytes (leading to the favourable outcome of HBsAg clearance), or can represent organ damage (with adverse outcomes including liver failure, long-term risk of fibrosis/cirrhosis, and death) 55 . Spontaneous flares of liver inflammation have been reported more commonly in males. For example, in a study of more than 1500 ethnically diverse adults followed over five years by the Hepatitis B Research Network in the US and Canada, flares were significantly commoner in males (OR 3.02; 95% CI: 1.59-5.74) 56 . Smaller CHB cohorts corroborate this finding, with more flares in males in a study of 217 asymptomatic HBeAg-negative patients (OR 4.5; 95% CI: 1.5-20.8) 57 and in 386 patients followed up in China to identify exacerbations 58 . In patients undergoing HBeAg seroconversion, male sex was again a significant risk factor for ALT flares 59 , and likewise in a study exploring flares following treatment withdrawal 60 . A prediction model for mortality in patients with acute on chronic liver failure in the setting of HBV infection includes male sex as a predisposing risk factor, with a view to informing improvements in stratified care 61 .
Based on existing data, it is not possible to determine whether hepatic flares in males are linked to immune control or to progressive organ damage, although given other evidence of enhanced pathology in males it is natural to hypothesise the latter. More data are needed to explore these observations, as liver flares are likely to be under-reported, and are clearly linked to other host and viral factors (HBV treatment, immunosuppression, HBV genotype, age, viral load (VL)) making it complex to disaggregate the overall influence of sex 55,56 .

Mechanistic influence of sex in liver disease associated with HBV infection
4.1 Immunological differences between males and females causing differential outcomes of infection Females typically have higher magnitude innate, humoral and cytotoxic responses in response to infection, broadly reflecting an immunostimulatory effect of oestrogen 6,7 , in contrast to the suppressive effect of androgens in males. Across a range of infectious diseases, this potentially explains the increased severity and/or duration of illness in males compared to females 62,63 , and enhanced vaccine responses in females (discussed in further detail in section 4.2 below). Sex dimorphism may be associated with differential expression of sex hormone receptors by immune cells (lymphocytes, monocytes and dendritic cells) and by hepatocytes. Chemokine and cytokine profiles also differ 64-66 , with androgens more typically associated with antiinflammatory cytokines. Sex-specific differences in IL-6 levels have been associated with HCC, with elevated levels in males leading to activation of AR, while oestrogen inhibits IL-6 production in females 67,68 .
Toll-like receptor 7 (TLR-7) is encoded on the X chromosome, one copy of which is inactivated in females to render TLR-7 expression equivalent to that in males. TLR-7 stimulation activates plasmacytoid dendritic cells, stimulating production of type I interferons and thus promoting T and B cell responses 69 , explaining why TLR-7 expression has been negatively correlated with HBV DNA VL 70 . Escape of the second copy of TLR7 from inactivation can increase TLR-7 expression in females 6,71 . Furthermore, a SNP in the TLR-7 gene (rs179008) in has been associated with protection from chronic infection in females (but not in males) 72 . Epigenetic modifications, and female X-chromosome mosaicism, may further contribute to this immunological advantage in females.

Sexual dimorphism in HBV vaccine responses
Males and females respond differently to HBV vaccination, with females mounting higher antibody titres than males 73,74 , and men more likely to be vaccine 'non-responders' (particularly in older age groups). For example, an Italian study reports that girls vaccinated after the age of 1 year mount a 1.2-fold higher median antibody titre than boys 73 . This is in line with trends reported for other childhood vaccines, with females producing increased antibody titres and longer term durability of response 74 . These effects may be due to the enhanced TLR-7 response in females, including a more inflammatory response to vaccine adjuvants. HBV vaccination in mice also demonstrates a higher rate of seroconversion in females, higher titres of anti-HBs antibody, higher magnitude T cell responses, and superior immunological memory 75 . It is uncertain to what extent these differences are significant to the role of the HBV vaccine in supporting elimination efforts, but they illustrate fundamental differences in the quality and quantity of the immune response, and may account for some degree of enhanced male susceptibility, although this is difficult to quantify. Despite the sex differences in vaccine response, vaccination may blunt the overall difference between males and females. Thus, newer studies of younger subjects, in which the prevalence of vaccine-mediated protection is high, may be less powered to detect sex differences compared to pre-vaccine cohorts.

Interaction between sex hormones and the HBV replication cycle
HBV has a complex replication cycle, reviewed elsewhere 76,77 . The virus has numerous interactions with host cell proteins, and by integration into the host genome can influence fundamental components of the cell cycle. The viral genome also has a direct influence through sex hormone response elements (previously reviewed 18 , and summarised in Figure 3). A pathway to oncogenesis in males occurs through involvement of the androgen receptor (AR) signalling pathway, which is common to cancer evolution caused by diverse viruses including HBV, EBV, HTLV-1, HHV8 and HPV 78 , and is also associated with the evolution of prostate cancer 79 . In HBV infection, AR stimulation is associated with increased expression of all four HBV mRNAs, via two androgen response elements (ARE) in the enhancer 1 (EnhI) region of the HBV genome 80 , leading to an upregulation in production of HBV proteins and DNA. In a mouse liver cell line, HB X protein increases AR activity in a dose-dependent manner 81,82 . Thus, a positive feedback loop operates between androgen exposure and X protein, which may explain the higher VL in males 18,83 (Figure 3). In human CHB infection, higher serum testosterone levels have been correlated with a higher incidence of HCC 84 , consistent with the male/female ratio in HCC. In mouse models, interleukin-6 (IL-6) is elevated in male HCC, leading to activation of signal transduction and activator of transcription 3 (STAT3), which upregulates AR 85 , and IL-6 has been suggested as a biomarker associated with worse outcomes in human HCC 86 .
In contrast to androgens, oestrogens may actively suppress HBV replication. Activation of the oestrogen receptor alpha (ER-α) can suppress HBV mRNA transcription by reducing expression of the transcription factor hepatocyte nuclear factor 4-alpha (HNF4-α). In a transgenic mouse model, HNF4-α increases HBV transcription by binding the enhancer I (EnhI) region of the HBV genome 87 . Deletion of ER-α in female mice removes the protective influence of oestrogen against hepatocarcinogenesis; relative to wild-type these knock-out animals had a nine-fold increased risk of tumorigenesis on exposure to carcinogens 88 . In cell culture experiments, estradiol reduces expression of the hepatocyte surface protein NTCP (sodium taurocholate co-transporting polypeptide), the main entry receptor for HBV 89,90 , which is another possible mechanism for protection in females 19 . Whether this mechanism is important in humans remains to be determined, but certain ER-α polymorphisms have also been shown to increase the risk of acute liver failure in patients with HBV 91 .

Sex to inform stratification for anti-viral therapy
Traditionally, HBV infection has been classified based on serology, VL, and markers of liver damage (such as liver enzyme levels, imaging scores and/or liver biopsy) 99 . Guidelines for nucleoside/nucleotide analogue (NA) therapy, based on these biomarkers together with age and sex, typically consider ~1 in 4 individuals with CHB to be treatment-eligible, although this varies between populations 100 . However, there is increasing recognition that current classification systems over-simplify liver disease by applying a traditional paradigm of linear progression, and that untreated individuals have a risk of progressive liver disease and HCC, and may also be at risk of transmitting infection 100-103 .
Female sex has been associated with a better response to PEG-IFN-alpha 104,105 , although this effect is not consistent across all studies 106 . Likewise, a female advantage is observed for nucleos(t)ide analogue treatment: in a study of >2000 HBV patients starting treatment with entecavir, earlier virologic response was observed in treatment-naïve females compared to males 107 . In HIV/HBV coinfected patients starting antiviral therapy, females have a higher rate of functional cure within two years (OR for HBsAg clearance in males 0.54 compared to females as reference group) 108 and male sex is reported in association with failure to suppress HBV viraemia 109 .
An increased understanding of the role of sex as a risk factor for disease could allow refinement of treatment algorithms, for example lowering the threshold for treatment of males and post-menopausal females, such that more individuals are treatment eligible. Treatment with NA agents such as tenofovir (typically prescribed as tenofovir disoproxil fumarate, TDF) is cheap, safe and effective in suppressing viraemia. Expanding CHB treatment is a key intervention to improve progress towards elimination goals 76,110 by offering therapy to those at greatest risk of long-term disease 111 as well as switching off transmission. Enhanced approaches to risk stratification will also benefit from an improved understanding of the interactions between sex and other attributes of the host (e.g. age, co-morbidity, gravidity, ethnicity) and virus (VL, genotype), which are yet to be fully resolved.
Sex and gender are likely to be significant factors in receipt of treatment. Universal health coverage (UHC) is included in SDG targets, setting a mandate for access to healthcare irrespective of demographic factors including sex, age, and ability to pay. For CHB, sex disaggregated data are currently insufficient to estimate the proportion of the untreated CHB population who are male vs. female 112 . However, treatment coverage may be influenced by biology and natural history of infection, in which higher VL and more advanced disease in males make them more likely to meet eligibility criteria than age-matched females.
Access to screening and engagement with care are also dependent on many other societal and infrastructure considerations, including education, beliefs and behaviour, and the structure of health-care services (discussed further in section 6). This topic has been explored for HIV, in which a higher proportion of women access treatment, have better immunological responses to treatment, and are retained in long-term follow-up 113,114 , but no such data are available for HBV infection.

Sex to inform HCC risk assessment and therapy
Risk scores have been proposed to predict HCC risk through measurement of non-invasive parameters such as VL, platelet count, and liver enzyme levels 115 , with risks of HCC increasing with VL >2000 IU/ml 116 . A recent systematic review and metanalysis evaluated the performance of these scores 117 . 12 of the 14 scores used sex as part of the algorithm, but the use of sex as a simple categorical variable may be over simplistic.
For example, the REACH-B score assigns a value of 2 to males and 0 to females 118 , without considering enhanced risks in older women. It should also be noted that there was a male predominance in the validation cohorts which may affect the assessment of relative risk between the sexes.
As evidence emerges for the role of sexual dimorphism in metabolic and malignant disease, this may shed light on new therapeutic approaches using hormonal therapy or blockade 119,120 . Androgen receptor (AR) blockade has thus far not shown therapeutic benefits for HCC 79,121,122 , but as we improve understanding of the biological pathways involved in driving cancer, this strategy warrants further exploration.

Interplay between age and sex
Alongside sex, another factor of considerable importance is age (first noted by Blumberg in his 1972 study, in which sexual dimorphism was most striking in the youngest groups 15 ). Given the strong observed effect of sex, it is to be expected that the natural history of CHB may change throughout life due to changes in the levels of sex steroids. For females, menarche, pregnancy, and menopause are therefore of particular relevance.
Earlier menarche has been correlated with earlier HBeAg seroconversion and a faster rate of HBsAg titre decline 123 . In contrast, early menarche has elsewhere been shown to correlate with an increased risk of HCC in HBsAg carriers 124 . A protective effect of early puberty has also been described in a small cohort of males 125 , perhaps counter-intuitively given that increased testosterone is otherwise identified as a risk factor. This reflects puberty as a period of complex immunological change, and highlights the need for further study. Postmenopausal status has been shown to reduce, or even remove, the protective effect of female sex. A multicentre cross-sectional study in China, involving 17,408 patients with CHB, found that the prevalence of cirrhosis increased at a faster rate after the age of 50 in females than in males 126 , and likewise sex differences diminished among older adults in a large Taiwanese cohort 28 . The risk of liver fibrosis in CHB has similarly been reported as comparable between post-menopausal females and age-matched males 127 . The loss of protection associated with menopause can be mitigated by hormone replacement therapy (HRT), which has a protective effect proportional to the duration of treatment 124 . Similar observations have been made in other chronic infections: in HCV infection, the risk of liver disease is lower in pre-menopausal females and accelerates to match that of males when the protective effect of oestrogen exposure is lost 128 ; in HIV the TLR-7 response and IFN production in women are dampened after the menopause, and restored by oestrogen replacement (reviewed in 129).
Together these data confirm that oestrogen has a protective and dose-dependent effect on the course and characteristics of CHB, which may vary depending on the life history of the individual female. The nuance of hormonal influence in the natural history of CHB in females requires more explicit study, but this is an area in which advanced understanding could impact on simple interventions such as enhanced surveillance, and antiviral and/or HRT prescription for post-menopausal women.

Maternal and child health
During pregnancy, there is a general tolerization of the immune system, with alterations in the Th1/Th2 ratio (with downregulation of Th1 immunity to avoid rejection of the foetus), such that Th1 cytokines (IFN-gamma and IL-2) are reduced, and there is a relative increase in Th2 cytokines (such as IL-4). These changes may have an impact on HBV infection in the mother 130,131 : during pregnancy, both increases and decreases in HBV DNA VL have been reported compared to non-pregnant women (reviewed in 132), while ALT flares are well recognised both during pregnancy and post-partum. Post-partum flares are mediated potentially by changes in the immune response (to re-set the non-pregnant Th1/Th2 balance), but can also be related to withdrawal of short-term antiviral therapy instituted to prevent vertical transmission 133-135 . In the majority of cases, flares are classified as mild (e.g. ALT up to 5 times upper limit of normal, ULN) or moderate (ALT up to 10 times ULN), with no clear detriment to maternal and foetal outcomes 136,137 , and indeed potentially associated with clearance of HBsAg. Rarely, severe flares (ALT >10 times ULN) arise, which can be associated with fulminant hepatitis 138 . Altered cytokine production and hepatic flares during pregnancy may stimulate clearance of HBeAg, and ultimately also loss of HBsAg (functional cure), although these events are uncommon 131 . In a Turkish cohort, multigravid females had a higher seroprevalence of HBsAg than primigravidae 139 , although this is difficult to interpret, as it may represent differences in exposure rates rather than a biological difference mediated by pregnancy. The risk of HCC in HBsAg positive women has also been inversely correlated with the number of pregnancies and the age of menopause 140 .
HBV infections in children typically occur either at birth (vertical transmission), or in the first few years of life through horizontal transmission from close household contacts 141 . The sex of the child may alter susceptibility, as evidenced by work in the HIV field, in which female fetuses have a 1.5-2-fold increased susceptibility to in utero infection compared to males 10 ; although there are no such data for HBV, this finding is a biological precedent for the impact of sexual dimorphism from the earliest days of life.
HBV vaccination of infants at birth is a simple and effective method of reducing early life infections, which account for most of the long-term burden of CHB. However, the effectiveness of birth vaccination is reduced when mothers have high HBV VL that is not treated during pregnancy 111 . Thus, although females typically have lower VL and a lower risk of long-term liver disease than males, provision of clinical care and interventions for women of child-bearing age and their infants (screening, monitoring, anti-viral therapy, and infant vaccination) are crucial public health interventions warranting sustained investment. The greater risk of persistent viral replication, cirrhosis, and HCC in individuals who have been vertically infected underlines the urgent need for diagnosis and intervention to prevent mother-to-child transmission and other early-life acquisition events 142 . Improved epidemiological data regarding HBV infections occurring in utero, in infancy and during childhood, stratified according to maternal characteristics and the sex of the offspring, could provide important insights into the biology of HBV transmission.
Route of transmission, HBV genotype, host genetics, and environmental factors may all interact with hormonal factors in mediating outcomes. For example, higher rates of vertical transmission in some settings (influenced by genotype) have been suggested to bring the male:female ratio closer to 1:1 in CHB 28 . Regional studies are particularly needed to provide an evidence base for high endemicity populations in the African and Asian subcontinents.

The role of gender
The construct of gender -defined as 'characteristics of women and men that are largely socially created' 143 -is important, accounting for social and behavioural factors. However, the term gender has often been used interchangeably with sex, and their influence can be difficult to disaggregate. The need to improving reporting of data representing both sex and gender is underpinned by Sex and Gender Equity in Research ('SAGER') guidelines, first published in 2016, to promote improved study design, analysis, and interpretation of clinical data 144 . However, progress in this domain has been slow, with little improvement in reporting sex-disaggregated data among organisations reviewed between 2018 and 2021 145 .
Gender roles may influence care-seeking behaviour, particularly through services for women, who typically have more points of contact with health services as a result of child-bearing and childcare, although traditional responsibilities are increasingly changing. Access to healthcare, education and support networks frequently vary by gender, from early in childhood and throughout life. Women are at increased risk of BBV transmission where their access to education is limited, and in the absence of reliable access to sexual and reproductive health services 146 . Access gaps may also be particularly pertinent for adolescent boys and young adult men; based on the HIV literature, these groups are the least likely to seek out or engage with diagnostic and treatment services 147,148 . For HBV, the long-term risk of complications is highest in these young men due to the combination of pathophysiology of infection in males, combined with the difficulty of providing consistent, accessible and acceptable healthcare. Men are also over-represented among people who inject drugs and the prison population, among whom risks of BBV exposure are higher 149 .
HBV is endemic in some communities of MSM and transgender women (e.g. in Papua New Guinea 150 and Nigeria 151 ), and providing reliable healthcare for these groups is challenging. Access to care for MSM, bisexual and transgender communities may be further complicated in countries where legislation prohibits same-sex relationships. A study in Spain to determine the incidence of acute HBV infection in people living with HIV reported that this was higher in males, particularly in MSM 151 , and HBV infection has also been associated with risky sexual practices and transactional sex 149,151,152 . High rates of drug and alcohol misuse increase the risk of exposure to infection with BBV, and accelerate liver disease in those with HBV infection, create barriers to interaction with healthcare services, and are associated with vaccine hesitancy in some populations 153 . Stigma is a barrier to diagnosis and intervention for HBV 154 , with discrimination potentially further enhanced by gender and sexuality. Extreme marginalisation in some of the groups at high risk of HBV infection is a significant barrier to education, diagnosis, treatment and prevention.

Conclusion
Sex and gender have a fundamental impact on the course and outcomes of HBV infection (Table 2), and developing insights in this domain has the potential to influence intervention ( Figure 4). The magnitude and nature of the sex effect vary between populations and with age, but -nearly forty years since the phenomenon of sexual dimorphism was first reported -have yet to be robustly quantified on a multiregional scale. Some regions and populations are yet to be represented in epidemiological data; at present there are insufficient data to permit meta-analysis representing the WHO African region. Improving the reporting of sex and gender data is imperative for HBV, as deepening an understanding of the biology and pathophysiology can inform new interventions, while stratification for treatment or surveillance by sex has the potential to improve outcomes for individuals, with associated population benefits in areas of high endemicity.
Females with CHB currently suffer from a lack of specific research, that potentially disadvantages HBV elimination efforts as a whole. While women are at lower risk of chronic infection and liver disease, a focus on women's health is nevertheless a fundamental aspiration for global health interventions for HBV, through reducing mother-to-child transmission, reducing the sex bias in data sets, and focusing attention on post-menopausal women who may be at increased risk of liver disease than younger females.  Enhancing an understanding of the mechanisms through which sex hormones mediate their influence can inform a better understanding of the pathophysiology of liver disease, with potentially important bearing on the use of existing interventions as well as informing the development of new therapies for HBV and the associated complication of HCC. Future research must focus on characterising the influence, impact and mechanisms of sexual dimorphism in HBV.

Data availability
No underlying data are associated with this article.

Describe Clearance Dynamics of Hepatitis B Surface Antigen (HBsAg) and
Open Peer Review In Figure 3, HNF4-α should be HNF4-α. The panel label (A, B, according to legend) is missing.

Is the review written in accessible language? Yes
Are the conclusions drawn appropriate in the context of the current research literature? Yes Reviewer comment: I have just two suggestions, which could supplement the current review. The studies presented in Table 1 do not include studies from Africa, and there is overall little information in the review in this context. It would be important to discuss this in more detail, particularly in the context of perinatal transmission where the reasons for sexual dimorphism might differ from the situation in sexual or blood-borne transmissions.

Reviewer comment:
The section on behavioural factors could be expanded with regard to sexual risk factors which could alter transmission rates, and factors which accelerate liver disease (e.g. alcohol and drug use). This topic is mentioned in section "6 The role of gender" but could be expanded given the potentially important role of behavioral factors. Reviewer comment: In figure 3 legend, hepatocyte nuclear factor 4-alpha abbreviation should be HNF4-α. Similar mistakes appeared in the text.
Author response: thank you for highlighting this; we have adjusted the figure legend and text accordingly. We have also corrected Erα to ER-α.
Competing Interests: No competing interests were disclosed.