Progress towards elimination of mother-to-child transmission of hepatitis B virus infection in China: a modelling analysis

Abstract

Objective

To determine the projected burden of hepatitis B virus (HBV) in China, the intervention strategies that can eliminate mother-to-child transmission (MTCT) by 2030 or earlier and the measurable parameters that can be used to monitor progress towards this target.

Methods

We developed a dynamic, sex- and age-stratified model of the HBV epidemic in China, calibrated using hepatitis B surface antigen (HBsAg) and e antigen (HBeAg) prevalence data from sequential national serosurveys (1979–2014) and the numbers of HBV-related cancer deaths (2012). We determined whether China can achieve elimination of MTCT of HBV by 2030 under current prevention interventions. We modelled various intervention scenarios to represent different coverage levels of birth-dose HBV vaccination, hepatitis B immunoglobulin to newborns of HBsAg-positive mothers and antiviral therapy (tenofovir) to HBeAg-positive pregnant women.

Findings

We project that, if current levels of prevention interventions are maintained, China will achieve the elimination target by 2029. By modelling various intervention scenarios, we found that this can be brought forward to 2025 by increasing coverage of birth-dose vaccination, or to 2024 by the administration of tenofovir to HBeAg-positive pregnant women. We found that achievement of the target by 2025 would be predicted by a measurement of less than 2% MTCT in 2020.

Conclusion

Our results highlight how high-quality national data can be combined with modelling in monitoring the elimination of MTCT of HBV. By demonstrating the impact of increased interventions on target achievement dates, we anticipate that other high-burden countries will be motivated to strengthen HBV prevention policies.

Introduction

The global momentum towards the elimination of viral hepatitis is growing. The World Health Assembly adopted the first global viral hepatitis elimination target in 2016 in calling for a 90% reduction in new chronic hepatitis B virus (HBV) infections by 2030, and the Global health sector strategy outlined a target of achieving a childhood prevalence of hepatitis B surface antigen (HBsAg) of 0.1% by 2030.¹

In China, the government have identified HBV as a significant health issue, and the virus is one of six major infectious diseases included in the latest 5-year Chinese national plan.² Substantial progress has been made over the past 20 years in implementing high national levels of coverage of interventions to prevent both horizontal transmission (infant HBV vaccination) and vertical transmission (birth-dose HBV vaccination within 24 hours of birth and hepatitis B immunoglobulin, Ig). The results of sequential national serosurveys in China reveal a reduction in the prevalence of HBsAg in children younger than 5 years from 9.7% in 1992, to 1.0% in 2006 and to 0.3% in 2014.³

Despite these public health efforts in China, there remains ongoing transmission of HBV. The major route is mother-to-child transmission (MTCT);⁴ even with the use of birth-dose HBV vaccination and hepatitis B Ig, about 2–9% of mothers who test positive for hepatitis B e antigen (HBeAg) transmit HBV to their newborns.^5,6 There is now accumulating evidence, particularly from China, that the additional use of tenofovir during the final trimester of pregnancy is a safe and effective method of further reducing MTCT.^7–10 However, the impact of this strategy at a population level is unknown.

In 2017, the National Health Commission of the People’s Republic of China (formerly the National Health and Family Planning Commission of China) and the World Health Organization (WHO) Western Pacific Region endorsed a framework for the triple elimination of MTCT of human immunodeficiency virus (HIV), HBV and syphilis by 2030.¹¹ The focus has now shifted from HBV control to the elimination of MTCT of HBV, the achievement of which by any country has not yet been validated by WHO. Elimination of MTCT is defined in China as an HBsAg prevalence of less than 0.1% in children aged 1–4 years, a directly measurable indicator in national serosurveys.

Given the global momentum and the available strategies to further reduce MTCT of HBV, there exists the real possibility of eliminating this route of transmission in China before 2030. Using a dynamic simulation model of the HBV epidemic in China, we aim to determine: (i) the current and projected burden of HBV nationwide; (ii) whether MTCT of HBV can be eliminated by 2030; (iii) which strategies could allow China to eliminate MTCT of HBV before 2030; and (iv) the indicators that can be used to measure progress towards this target.

Methods

Model parameterization

We developed a dynamic, age- and sex-stratified simulation model of the HBV epidemic in China at a national level, incorporating the latest locally available epidemiological and demographic data. We based our dynamic model on a previously published transmission model that has been parameterized and calibrated to China-specific data (Table 1).²⁰ Our model is calibrated to HBsAg prevalence from four large nationally representative serosurveys (conducted in 1979, 1992, 2006 and 2014; available in the data repository),^3,13,17,21 HBeAg prevalence from 2006 and 2014, and the numbers of HBV-related cancer deaths in China (using overall liver cancer deaths²² with a 70.0% HBV-attributable fraction applied).²³ We stratified MTCT percentages and the efficacy of prevention of MTCT (PMTCT) interventions by HBeAg status, and parameterized these with data from China (data repository;²¹ Table 1). Regarding the efficacy of combined birth-dose HBV vaccination, hepatitis B Ig and tenofovir treatment for HBeAg-positive mothers, we assumed a residual transmission of 1.0%, consistent with an 80.0% transmission reduction compared with birth-dose vaccination and hepatitis B Ig only.⁹

Table 1. Sources of transmission, calibration and intervention coverage data used to calibrate dynamic modelling of hepatitis B virus, China, 2019 .

Variable	Value (%)	Source
MTCT parameters
HBeAg− mother, birth-dose vaccination	0	Lu et al.12
HBeAg+ mother, birth-dose vaccination	0.127	Ying et al. (2017, personal communication)
HBeAg− mother, birth-dose vaccination and hepatitis B Ig	0	Lu et al.12
HBeAg+ mother, birth-dose vaccination and hepatitis B Ig	0.056	Ying et al. (2017, personal communication)
HBeAg+ mother, birth-dose HBV vaccination, hepatitis B Ig and tenofovir	0.01	Hyun et al.9 (to represent an 80% reduction in MTCT with addition of tenofovir)
Calibration data
Prevalence of HBsAg by sex and age, years
Male, 1–4	9.71	China CDC National Serosurvey 197913
Male, 5–14	11.45–11.96
Male, 15–59	7.38–11.55
Male, 60–89	1.29–5.06
Female, 1–4	7.92
Female, 5–14	7.79–9.08
Female, 15–59	5.72–7.76
Female, 60–89	3.66–4.55
Male, 1–4	9.18	China CDC National Serosurvey 199214
Male, 5–14	10.89–11.94
Male, 15–59	8.07–12.28
Male, 60–64	11.76
Female, 1–4	7.30
Female, 5–14	7.88–9.07
Female, 15–59	5.87–8.97
Female, 60–64	14.81
Male, 1–4	1.21	China CDC National Serosurvey 200615
Male, 5–14	1.82–3.77
Male, 15–59	8.48–10.91
Female, 1–4	0. 93
Female, 5–14	1.37–2.95
Female, 15–59	5.79–7.47
Male, 1–4	0.35	China CDC National Serosurvey 201416
Male, 5–14	0.47–1.28
Male, 15–29	2.82–6.32
Female, 1–4	0.28
Female, 5–14	0.70–1.43
Female, 15–29	2.70–4.88
Prevalence of HBeAg by sex and age, years
Male, 1–4	70.75	China CDC National Serosurvey 200617
Male, 5–14	66.38–70.54
Male, 15–59	10.60–52.67
Female, 1–4	66.20
Female, 5–14	64.71–68.35
Female, 15–59	11.86–56.79
Male, 1–4	87.46	China CDC National Serosurvey 20143
Male, 5–14	53.42–66.47
Male, 15–29	37.03–59.38
Female, 1–4	90.17
Female, 5–14	32.97–61.37
Female, 15–29	15.62–44.70
Intervention coverage		Liu et al.,18 WHO19
Infant vaccination, 1992–2018	32.6–99.49
Birth-dose vaccination,1992–2018	23.5–96.05

−: negative; +: positive; CDC: Centers for Disease Control and Prevention; HBeAg: hepatitis B e antigen; HBsAg: hepatitis B surface antigen; HBV: hepatitis B virus; Ig: immunoglobulin; MTCT: mother-to-child transmission; WHO: World Health Organization.

We parameterized the historical coverage of infant and birth-dose vaccination by using annual national immunization programme data.¹⁸ Hepatitis B Ig administration became national policy in 2010, and coverage was 91.2% in 2013 and 99.5% in 2016. The use of tenofovir by pregnant women with a high viral load (> 6–7 log₁₀ international units per mL) to prevent MTCT of HBV is recommended by the Asian Pacific Association for the Study of the Liver 2015 guidelines,²⁴ but this intervention is usually limited to tertiary health-care centres and is not national policy. We assumed that historical population-level use of tenofovir during pregnancy was zero.

In all analyses, our primary outcome measure was whether China can achieve elimination of MTCT of HBV and the year of achieving that target. Our secondary outcome measure was the cumulative number of new infections averted under the different intervention scenarios.

Intervention scenarios

We first modelled the epidemic under a status quo scenario, that is, intervention coverage remaining at current (2016) levels. We then modelled different scenarios to represent the various increases in and additions to this status quo intervention from 2019 (Table 2). In Scenario A, birth-dose vaccination is increased from 95.9% in 2016 to 99.0%. In Scenario B, in addition to the birth-dose vaccination increase in Scenario A, hepatitis B Ig is also increased from 99.5% in 2016 to 99.9%. Scenarios C–E represent Scenario B with the further addition of peripartum antiviral treatment, namely the administration of tenofovir to 50.0%, 70.0% and 90.0% of HBeAg-positive pregnant women, respectively. We used HBeAg status as a proxy for high viral load, as this is supported by data from China.¹²

Table 2. Main intervention scenarios considered in achieving the elimination of mother-to-child transmission of hepatitis B virus, China, from 2019.

Intervention scenario	% of relevant population covered
	Birth-dose HBV vaccination of all newborns	Hepatitis B Ig to newborns of HBsAg+ mothers	Tenofovir to HBeAg+ pregnant women
Status quo (2016)	95.9	99.5	0.0
A: Increased birth-dose vaccination from 2019	99.0	99.5	0.0
B: Scenario A with increased hepatitis B Ig from 2019	99.0	99.9	0.0
C: Scenario B with tenofovir	99.0	99.9	50.0
D: Scenario B with tenofovir	99.0	99.9	70.0
E: Scenario B with tenofovir	99.0	99.9	90.0

+: positive; HBeAg: hepatitis B e antigen; HBsAg: hepatitis B surface antigen; HBV: hepatitis B virus; Ig: immunoglobulin.

Note: Infant vaccinations remain unchanged from status quo in all scenarios.

Further analysis

Although elimination of MTCT of HBV is defined in China as a prevalence of HBsAg of less than 0.1% in children aged 1–4 years, there are alternative age ranges that can be used. We therefore also modelled the outcome of status quo interventions when changing the age group in which prevalence is measured to all children younger than 5 years and to children exactly 5 years.

We also performed a sensitivity analysis to explore how five different fertility projections would affect outcomes. These projections included three United Nations estimates of fertility at median, high and low levels,²⁵ and two hypothetical scenarios representing the relaxation of China’s one-child policy (i.e. increased fertility) where fertility returned to 1965 levels for women (i) of all ages or (ii) older than 30 years.

Separately, we also evaluated the impact of (i) reducing HBV prevention interventions to 50.0% and 80.0% of current levels; and (ii) catch-up vaccination campaigns for children aged 2 years, 2–15 years and 12–13 years, in which 100.0% of unvaccinated and susceptible children in each age group were vaccinated (data repository).²¹

Finally, we also determined the cost of an intervention package that would allow a population strategy of peripartum tenofovir to be cost-effective. This cost represents the package required for the successful delivery of the intervention, and would include antiviral drugs for 4 months, diagnostics needed to select those in need of treatment (HBeAg status as a proxy for HBV viral load) and treatment monitoring costs. We used a cost–effectiveness threshold of United States dollars (US$) 4062 per disability-adjusted life year averted, which represents half the gross domestic product per capita of China, in line with current recommended international guidelines.²⁶

Measuring progress

To monitor progress towards achieving elimination of MTCT of HBV, we considered the relation between MTCT values and the year in which the elimination prevalence target would be reached for both (i) all HBsAg-positive mothers and (ii) HBeAg-positive mothers only. For each of our five intervention scenarios, we plotted the projected MTCT in 2020 (the intervention having been scaled up in 2019) against the corresponding year of achievement of the elimination target.

Results

HBV burden

Our model estimates that in 2017 there were 78 million people living with chronic HBV infection, consistent with existing burden estimates.^27,28 At current levels of coverage, we project 198 000 new chronic infections and 7.7 million deaths between 2019 and 2030 (data repository).²¹

Status quo

By maintaining current intervention programmes, China is predicted to reach the elimination target of less than 0.1% HBsAg prevalence in children aged 1–4 years by 2029 (Fig. 1). This reduction in the epidemic at status quo levels of intervention is explained by the fact that, with time, the year in which a woman of child-bearing age was born corresponds to higher vaccination coverage levels at birth, and therefore lower HBsAg prevalence among pregnant women in future years (data repository).²¹

Fig. 1.

Projections of prevalence of hepatitis B surface antigen among children aged 1–4 years for current levels of interventions to prevent transmission of hepatitis B virus, China, 2015–2050

HBsAg: hepatitis B surface antigen.

Intervention scenarios

Compared with the status quo, our model shows that by increasing coverage of birth-dose HBV vaccination from 95.9% in 2016 to 99.0% (Scenario A), 54 000 new chronic infections will be avoided by 2030 (Fig. 2) and the elimination target will be reached by 2025 (Fig. 3). The incremental impact of an increase in hepatitis B Ig coverage (Scenario B) is small as coverage is already very high, and does not change the year of target achievement (Fig. 4). The addition of tenofovir administration to pregnant women at 50.0%, 70.0% and 90.0% coverage levels (scenarios C, D and E, respectively) would avoid a further 37 000, 52 000 and 67 000 new chronic infections, respectively, compared with Scenario B (Fig. 2), and bring the year of elimination target achievement forward to 2024 in all cases (Fig. 4).

Fig. 2.

Impact of intervention scenarios to prevent mother-to-child transmission of hepatitis B virus on the number of cases averted, China, 2015–2050

Note: As in Table 2, scenarios are defined as increased coverage of birth-dose vaccination (A), plus increased coverage of hepatitis B Ig (B), plus 50.0% (C), 70.0% (D) and 90.0% (E) coverage of tenofovir.

Fig. 3.

Impact of intervention scenarios to prevent mother-to-child transmission of hepatitis B virus on the prevalence of hepatitis B surface antigen among children aged 1–4 years, China, 2015–2050

HBsAg: hepatitis B surface antigen.

Note: As in Table 2, scenarios are defined as increased coverage of birth-dose vaccination (A), plus increased coverage of hepatitis B Ig (B), plus 50.0% (C), 70.0% (D) and 90.0% (E) coverage of tenofovir.

Fig. 4.

Impact of intervention scenarios to prevent mother-to-child transmission of hepatitis B virus on year of target achievement, China

Note: The target is a prevalence of hepatitis B antigen less than 0.1% among children aged 1–4 years. As in Table 2, scenarios are defined as increased coverage of birth-dose vaccination (A), plus increased coverage of hepatitis B Ig (B), plus 50.0% (C), 70.0% (D) and 90.0% (E) coverage of tenofovir.

Further analysis

We show in data repository how the year of achieving the elimination target changes with the definition of elimination; if HBsAg prevalence of less than 0.1% in children aged 0–5 years or exactly 5 years is the target, this would be achieved in 2028 and 2030, respectively.²¹

Our different hypothetical fertility projections show that, if fertility levels return to pre-1965 levels in women of all ages or only in women older than 30 years, the projected fall in prevalence of HBsAg in children will slow. This reduced rate of prevalence decline would mean that the year of achieving the elimination target would be delayed to 2030 and 2032, respectively (data repository).²¹ This result is due to the fact that HBsAg prevalence among older women is higher, as they are less likely to have benefitted from the vaccination programme.

We show that a reduction in prevention interventions will slow down or even reverse the decline in HBsAg prevalence (data repository).²¹ A reduction in coverage of birth-dose HBV vaccination to 80.0% or 50.0% of current levels would lead to 278 000 and 803 000 further new infections, respectively, during 2019–2030, and would push back the date of elimination target achievement to 2036 and 2040, respectively.

We project that catch-up vaccination campaigns will have a lower impact relative to the five main intervention scenarios. We estimate that 1800 new chronic infections will be averted by 2030 by targeting children aged 2–15 years, but this number is less than 500 if groups aged 2 or 12–13 years are targeted. We also project that catch-up programmes will not bring the year of elimination target achievement any closer (data repository).²¹

We estimate that administration of peripartum tenofovir would be cost-effective if the price of an intervention package falls between US$ 11 (for a 10% discount rate) and US$ 197 (for a 3% discount rate), assuming a long-term perspective.²⁹ However, if a population-level scale-up of antiviral treatment for those with chronic HBV infection became policy, the cost of the intervention package would need to be even lower to be cost-effective.

Measuring progress

We show how measured MTCT in a particular year can be used to predict the year in which the elimination target will be achieved. For example, if in 2020 we measure MTCT values of less than 2.1% among all HBsAg-positive mothers (Fig. 5) and of 6.5% in only HBeAg-positive mothers (data repository),²¹ we can predict that the elimination target will be achieved by 2025.

Fig. 5.

Relationship between measured mother-to-child transmission in mothers positive for hepatitis B surface antigen and the year in which the elimination target will be achieved, China

HBsAg: hepatitis B surface antigen; MTCT: mother-to-child transmission.

Note: (i) As in Table 2, scenarios are defined as increased coverage of birth-dose vaccination (A), plus increased coverage of hepatitis B Ig (B), plus 50.0% (C), 70.0% (D) and 90.0% (E) coverage of tenofovir. (ii) Triangles represent estimated mother-to-child transmission of hepatitis B in 2020 for intervention scenarios A–E.

Discussion

Our modelling study has shown that, at the current high levels of prevention interventions, China could be one of the high-burden countries to achieve elimination of MTCT of HBV before the 2030 global targets. However, our results also highlight that these high levels of prevention interventions must be maintained, as any reductions in coverage will delay the achievement of the elimination target.

Although we have also demonstrated that the addition of tenofovir treatment in pregnancy as an extra method of preventing MTCT will have a substantial impact on the trajectory of the epidemic in China, it must be noted that this result relies on certain assumptions about the population-based efficacy of tenofovir treatment. Our analysis provides projections of the possible impact under ideal conditions, including high levels of coverage and adherence to a peripartum antiviral therapy strategy. However, whether these conditions can be achieved in a real-world setting will also depend on the availability of good-quality health-care services, adherence of patients to HBV antiviral therapy during pregnancy and the availability of screening tests with high diagnostic performance. Our projections are also dependent on how the Chinese government can increase these prevention strategies by such a large scale, overcoming logistical barriers and increasing education among health-care workers and patients regarding the importance of PMTCT of HBV strategies. Implementation projects such as the Shield Project, which are piloting methods of increasing such interventions in China,⁴ will provide much-needed real-world data on MTCT of HBV and on innovations to reduce vertical transmission in China. Such empirical data will help to refine our model projections, and can become part of the iterative process where applied modelling is used to help with programmatic support.

By using HBeAg status (an accurate marker of viral load, as confirmed in a recent global systematic review)³⁰ to determine which pregnant mothers require antiviral therapy, we also address the issue of limited access to HBV deoxyribonucleic acid testing in some areas of China. Further research into the most cost-effective method of scaling up tenofovir administration to pregnant women is needed, and the approach taken could be tailored by rural or urban areas depending on the availability of HBeAg or HBV viral load testing. Moving away from a one-size-fits-all solution in China might overcome differential logistical and financial barriers.

Over the last few years, the price of tenofovir has been successfully negotiated in China and has fallen dramatically from nearly US$ 1000 per patient per year in 2016 to US$ 10 per patient per year in 2020 (i.e. the drug cost alone meets our threshold for cost–effectiveness). However, although the cost of the antiviral drug alone should no longer be a barrier, the financing of a large-scale intervention providing tenofovir to pregnant women will still need careful consideration as there will also be associated diagnostic and monitoring costs.

China is exceptional in having managed to achieve such high levels of hepatitis B Ig coverage to babies born to all HBsAg-positive mothers. In many low- and middle-income countries, the use of hepatitis B Ig is limited by cost, lack of availability, cold-chain requirements and the high numbers of births outside hospitals.³¹ Data are emerging that show there is no residual transmission in HBeAg-negative mothers if a birth-dose HBV vaccination is received;¹² providing hepatitis B Ig only to babies born to HBeAg-positive mothers could therefore be a cheaper strategy with a potentially similar impact. Another strategy that might be useful for other countries with financial and logistical constraints to scaling up hepatitis B Ig is to adopt a PMTCT intervention based on birth-dose vaccination and tenofovir only. However, the efficacy of such a strategy is unknown; trials are currently ongoing in Lao People's Democratic Republic and Thailand.³²

Our study benefitted from projections that used a dynamic model calibrated to the latest national data in China. Although previous models have attempted to quantify the HBV epidemic in China and the impact of vaccination,^33,34 they were not calibrated to the latest available data, did not evaluate the impact of the addition of hepatitis B Ig or tenofovir, or were not dynamic in nature. A further strength of our study is our proposal of a novel method to monitor HBV elimination targets both in the Western Pacific Region and globally. Our method, which combines measurable indicators on programmatic coverage and measured MTCT values with modelling, could overcome the reliance on expensive serosurveys that would otherwise be required. Antenatal screening combined with routine post-vaccination serological testing in babies born to HBsAg-positive mothers could be a useful method of measuring MTCT percentages and monitoring the effectiveness of intervention strategies. However, further research is needed on optimal and cost-effective methods of post-vaccination serological testing, and how these could be integrated into existing practices without overburdening health systems.

Our study had some limitations. We have taken a national perspective using national programme and serosurvey data; however, this may obscure heterogeneity on a more provincial level and does not consider the effect of migration or HIV coinfection on the projections (although the effect of the latter is expected to be small in China). Furthermore, we did not investigate the costs or the cost–effectiveness of alternative methods of implementing such a programme, or how non-adherence to tenofovir therapy during pregnancy would affect the trajectory towards elimination. All of these factors are important to policy-makers in deciding whether such a strategy should be adopted.

Our study highlights the important role of dynamic modelling combined with high-quality national data in validating country-level elimination of MTCT of HBV. To monitor progress on this path to elimination, we have proposed a unique framework for measurable indicators. Our results should motivate other high-burden countries that have not yet attained such high levels of PMTCT intervention coverage to not only strengthen their HBV prevention strategies, but also consider the provision of tenofovir to pregnant HBeAg-positive women and the establishment of effective monitoring systems.