Abstract

Background

Hepatitis B virus (HBV) surface antigen (HBsAg) seroprevalence was high before the national vaccine policy was introduced in Taiwan, indicating significant HBV infection rates. The success of the HBV immunization program and other preventive measures likely led to decreased HBsAg prevalence among pregnant women. This study reports on the HBV seroprevalence among pregnant women in Taiwan from 2016 to 2021, including those potentially affected by the universal hepatitis B vaccination at birth.

Methods

This claim-based cohort study included pregnant women with hospital-based prenatal HBV screening data: 162,662 for HBsAg and 161,729 for HBeAg, from 2016 to 2021. Patient medical records were reviewed to collect information on demographic characteristics and other health conditions. Logistic regression models were used to identify risk factors associated with HBsAg and HBV e antigen (HBeAg) positivity.

Results

The seroprevalence for HBsAg and HBeAg during the study period was 4.0% and 0.6%, respectively. HBsAg positivity was highest among women born before July 1984 (pre-vaccination period; 8.6%), decreasing to 2.2% among those born between July 1986 and 1988 (national vaccination implementation) and further declining to 1.1% for those born after 1997. These data underscore the crucial role of large-scale immunization strategies in controlling HBV infections. Similarly, HBeAg positivity was highest among pregnant women born before the vaccination program (~ 1.0%), decreasing significantly to 0.4% for those born after 1989. The results showed geographic variations, potentially reflecting factors such as the mother’s age and foreign nationality. However, the birth year was the most crucial factor associated with HBV marker positivity.

Conclusions

The implementation of national vaccination programs has demonstrated significant success in reducing HBV seroprevalence among pregnant women, which is particularly evident in the substantial decrease in HBsAg seroprevalence in Taiwan post-July 1986. These findings emphasize the importance of continued and consistent vaccination efforts, supporting the need for ongoing public health strategies to combat HBV infections effectively.

Background

Despite the adoption of effective vaccines and antiviral treatments over the past decades, the hepatitis B virus (HBV) continues to pose a significant global public health challenge [1,2,3]. In 2016, the World Health Assembly declared its objective to eliminate all viral hepatitis by 2030 [4]. Nonetheless, as of 2019, the World Health Organization (WHO) reported a staggering 820,000 deaths attributed to HBV infection, with 296 million individuals enduring chronic HBV infections [5]. Notably, mother-to-child transmission (MTCT) continues to be the predominant cause of new HBV infections, accounting for an estimated 1.5 million cases among neonates [6].

Based on various studies, global HBV endemicity falls within the intermediate range (HBV surface antigen [HBsAg] seroprevalence = 3.5%–5.6%) [1] across all ages, indicating that approximately 300–400 million individuals live with HBV infections worldwide. Encouragingly, the HBV seroprevalence is significantly lower in younger age groups, with a global estimate of 1.3%–3.4% for children aged < 5Ìýyears [1]. Before the implementation of the HBV vaccination program in Taiwan, the country faced some of the highest HBV rates globally. However, since the launch of the world’s first universal HBV vaccination program for infants in Taiwan 30Ìýyears ago, significant progress has been made. The HBV carrier rate among children under 6Ìýyears old has dramatically decreased from 10.5% to less than 0.8% [7]. To achieve the WHO goal of a 95% reduction in chronic HBV infections by 2030, the HBsAg seroprevalence should be ≤ 0.1% among children aged ≤ 5Ìýyears [8].

The transmission of HBV occurs through contact with blood and bodily fluids. Despite vaccination efforts, approximately 10% of pregnant women with a high hepatitis B viral load remain unable to prevent vertical transmission of the virus to their offspring. Government-funded prenatal exams in Taiwan currently include HBsAg and HBV e antigen (HBeAg) testing for nearly all pregnant women. The timing of testing was moved forward from the fifth (in the third trimester) to the first (before 12Ìýweeks of gestational age) prenatal exam in November 2014 [9]. Prenatal screening data from the National Immunization Information System showed that seropositivity for HBsAg and HBeAg was 5.9% and 1.0% in 2016, respectively [9]. While several studies have demonstrated a significant reduction in HBsAg seroprevalence following the initiation of Taiwan’s national HBV vaccination program, there is a notable gap in national-level data beyond 2016. [10] Moreover, many studies have relied on estimates from small populations [11].

Since 2017, Taiwan’s National Health Insurance (NHI) has provided researchers with access to a claim-based research database (NHIRD) containing laboratory data [12]. Therefore, this study retrospectively analyzed the NHIRD to assess the real-world burden of HBV among pregnant women in Taiwan who have undergone routine HBsAg testing-an essential indicator for MTCT between 2016 and 2021.

Methods

Data source, security, and quality control

This retrospective cohort study used the NHIRD in Taiwan, which incorporates data from over 20,000 contracted healthcare facilities, covering 92% of nationwide hospitals and clinics and representing more than 99% of the Taiwanese population.

Data protection and permission

Access to the NHIRD data was obtained for research purposes, following strict privacy guidelines. Personal identifiers were systemically removed to safeguard patient anonymity during the analysis. The data processing and analysis adhered to NHI Administration (NHIA) regulations and were performed using the Applied Health Research Data Integration Service provided by the NHIA, ensuring meticulous data management and transformation.

Data selection and definitions

This study focused on pregnant women aged 10–60Ìýyears undergoing prenatal care visits for HBV testing in Taiwan between 2016 and 2021. Although an initial cohort size of approximately 1,100,000 pregnant women was anticipated, those who underwent prenatal HBV screening at clinics were excluded since less than 0.1% of laboratory data are reported in these settings. The data specifically extracted for analysis were from the first prenatal examination, typically conducted before 12Ìýweeks of gestational age, to measure HBV seroprevalence. This study ultimately included around 160,000 individuals with valid HBV test reports within 30Ìýdays of testing. FigureÌý1 illustrates the detailed sample selection process. The main outcomes of interest were HBsAg and HBeAg.

Fig.Ìý1

Participant selection process

Study variables

The individual-level variables analyzed included age at pregnancy, socioeconomic status, geographic distribution, comorbidities (HIV, drug abuse, and HCV), medication use (anti-HBV and anti-HCV therapies), and hepatitis screening tests conducted during the 1-year baseline period. Detailed diagnostic codes for comorbidities, medication usage, and hepatitis screening are presented in Table S1. The examined provider-level variables were the facility type (medical center, regional or local hospital) and geographic location (seven regional living circles) of the HBV testing sites. Due to inconsistencies in vaccination policy, the nationality of pregnant women was found to be associated with HBsAg seroprevalence [13]. However, due to limitations in obtaining individual-level data, this study chose to use county-level data on the prevalence of mothers with foreign nationality, which ranged from 1.71% to 6.60% during the study period, to assess its impact on HBsAg seroprevalence.

Statistical analysis

The primary findings were the seroprevalence of HBsAg and HBeAg among pregnant women during this specific period. The secondary findings included a comparison of seroprevalence by age at pregnancy and other sociodemographic variables. The baseline characteristics of pregnant women who underwent HBsAg and HBeAg testing are presented, with numerical variables described using means with standard deviations and categorical variables presented as frequencies and percentages. Groups were compared using standardized mean difference or Chi-square tests. Factors associated with HBsAg or HBeAg positivity were identified through stepwise logistic regression, considering a p-value < 0.2. The statistical analyses were performed using SAS/STAT 9.4 (SAS Institute, Cary, NC, USA).

Results

Study cohort

Of the 1,104,149 pregnant women tested for HBV between 2016 and 2021, 53.2% (n = 587,674) were tested in hospitals, 46.6% (n = 515,131) belonged to the pre-national HBV immunization program birth cohort, 61.1% (n = 674,602) fell within the middle-income range (20,000–39,999 NTD per month; 1 USD ≈ 31 NTD; based on the average exchange rate of NTD against USD between January 2016 and December 2021), and around 50% (n = 526,965) resided in Northern Taiwan. Regarding testing facility type, 10.1% were tested in medical centers, 16.5% in regional hospitals, 26.6% in local hospitals, and 46.8% in clinics (TableÌý1). We compared baseline characteristics by facility type. A higher percentage of pregnant women tested in medical centers were born before June 1986. Also, those receiving HBV tests in medical centers had higher monthly incomes than those tested at other facilities (Additional file 1: Table S2, Figure S1).

TableÌý1 Baseline characteristics of all pregnant women undergoing HBV testing

Among pregnant women with HBV test reports, only 15% had laboratory data available (n = 162,689 for HBsAg and n = 162,615 for HBeAg; Additional file 1: Table S3 and Table S4). The detailed cleaning process for HBsAg and HBeAg laboratory data is shown in Additional file 1: Figure S2. Fortunately, no significant differences in baseline characteristics existed between those with and without laboratory data, and > 99% of results could be classified as positive or negative based on the available reports.

Seroprevalence for HBsAg and HBeAg by birth cohort

Throughout the study period, the overall seroprevalence for HBsAg and HBeAg was 4.0% and 0.6%, respectively (Additional file 1: Table S5 and S6). Age-related variations in HBsAg and HBeAg seroprevalence were observed across birth cohorts. Notable trends included a peak in seroprevalence of 8.6% for those born before June 1984, decreasing to 4.0% for those born between July 1984 and June 1986, a slight decrease to 2.2% for those born after June 1986, and a significant decrease to 1.1% for those born after 1997. These findings inform targeted public health strategies and underscore vaccination program effectiveness. In contrast, HBeAg seroprevalence varied slightly over the study period, ranging from 0.99% to 0.40% (Fig.Ìý2).

Fig.Ìý2

Seroprevalence of HBsAg and HBeAg positively by birth cohort

Seroprevalence for HBsAg and HBeAg by regions

HBsAg and HBeAg seroprevalence exhibited distinct patterns across demographic factors. HBsAg positivity was highest among women with the highest monthly income (4.3%) and those tested in medical centers (4.3%). Regional variations included a highest seroprevalence of 4.6% in Yunlin, Chiayi, and Tainan. Foreign nationality was positively associated with HBsAg positivity (OR = 1.20, 95% CI 1.12–1.28 for high level of mother’s foreign nationality). HBeAg positivity was highest among those with the lowest income level (0.7%), using local hospitals (0.7%), and using facilities in outlying islands (1.3%; Additional file 1: Table S5 and S6).

FiguresÌý3 present the age-unadjusted and -adjusted HBsAg seroprevalence across seven regional living circles. The unadjusted HBsAg seroprevalence was highest in Yunlin, Chiayi, and Tainan (4.62%) and lowest in Penghu, Kinmen, and Lienchiang (3.45%). After age adjustment, Yunlin, Chiayi, and Tainan still had the highest HBsAg seroprevalence (4.46%), and Penghu, Kinmen, and Lienchiang still had the lowest (3.78%). These findings underscore persistent regional disparities in HBsAg seroprevalence after adjusting for age differences.

Fig.Ìý3

Age-unadjusted and age-adjusted seroprevalence for HBsAg in seven regional living circles. The dark-to-light color gradient represents the highest to lowest seroprevalence

Assessing HBeAg seroprevalence using the same approach, the unadjusted seroprevalence was highest in Penghu, Kinmen, and Lienchiang (0.94%) and lowest in Taipei City, New Taipei City, Keelung, and Ilan (0.48%). This pattern persisted after adjusting for age. However, due to a small sample size in some counties, we do not report HBV seroprevalence by county, ensuring the reliability and accuracy of our findings.

Factors associated with HBsAg positivity

Table 2 shows the factors associated with HBsAg positivity. Crude ORs indicated associations with screening year, birth cohort, hospital level, regional living circles, mother’s foreign nationality, comorbidities, medication use, and hepatitis screening. However, the small sample size in the medication group resulted in wider 95% CI, reflecting greater uncertainty in the estimation. Adjusted ORs were determined for variables with a p-value < 0.2 (data not shown). Ultimately, only birth cohort was significantly associated with HBsAg positivity. Compared to those born before June 1984, adjusted ORs revealed reduced risks for those born between July 1984 and June 1986 (0.43, 95% CI: 0.40–0.47), between July 1986 and 1988 (0.23, 95% CI: 0.24–0.25), and after 1989 (0.17, 95% CI: 0.16–0.18). We do not report the regression analysis for HBeAg due to poor model performance and logistic regression convergence issues.

TableÌý2 Stepwise logistic regression analysis of factors associated with HBsAg positivity

Discussion

This study investigated HBV prevalence among pregnant women by analyzing nationwide data on HBsAg and HBeAg seroprevalence in Taiwan. Our analyses revealed a significant decline in HBsAg seroprevalence following the introduction of the national HBV vaccination program. Despite this decline, HBsAg seroprevalence remained around 2%, which could be attributed to factors such as immigration from countries without universal infant vaccination programs or immunoprophylaxis failure. The precision of official HBV prevalence data may be influenced by constraints such as incomplete information about immigrants or inaccuracies in government databases. These limitations could potentially cause the impact of the national vaccination program to be underestimated, underscoring the importance of ongoing attention to data collection and reporting practices.

The nationwide HBV immunization program, launched in Taiwan in July 1984, has proven highly successful in reducing HBsAg seroprevalence [6], which markedly decreased from 10.5% [14] before the immunization program to 0.78% [15] in 2005 among children aged 6–7 years [16]. Long-term follow-ups conducted 30 years post-implementation revealed that HBsAg seroprevalence had decreased below 2% for those born after 1986, falling below 1% for those born after 1992 [17].

In Taiwan, the current immunization schedule for HBV vaccines in infants involves doses administered within 24 h of birth, at 1 month, and at 6 months. This strategy aligns with similar initiatives inÌýHong Kong, China, andÌýKorea [18,19,20], where HBV vaccines are administered to newborns, followed by booster doses at one and six months. Despite varying programs, these Asian countries have met the WHO’s interim target of a 1% prevalence of chronic HBV infections among children aged five years in 2020 [18, 21, 22]. This observation underscores the effectiveness of immunization programs in reducing HBsAg seroprevalence, emphasizing the need for their continual promotion and improvement to further alleviate the HBV burden and safeguard the population from this infectious disease.

The HBV prevalence reported based on a national survey in the US was deemed an underestimate due to the under-sampling of immigrants; the projected HBV prevalence of 0.55% was significantly higher than the nationally reported estimate of 0.32% [23]. However, the HBV seroprevalence in our study is presumed to reflect the actual national prevalence in Taiwan. This assertion is based on data derived from the NHI, which covers over 99% of residents in Taiwan regardless of nationality. In Taiwan, 60% of foreign spouses were from China, with approximately 20% and 5% coming from Vietnam and Indonesia, respectively [10]. These regions exhibit high intermediate HBV endemicity. The HBsAg seroprevalence of the general population in China decreased from 9.6% in 1973–1984 to 3.0% in 2021 [24]. Similarly, the HBV prevalence was 8.1% in Vietnam in 2017 [25] and 7.1% in Indonesia in 2013 [26]. Our findings also indicate a higher prevalence in regions with more foreign mothers. As our study lacked information on nationality, it cannot exclude those not born in Taiwan. The high proportion of foreign spouses in Taiwan from these countries might explain why the HBV rate is still above 0.1% in Taiwan, potentially resulting in an underestimation of the impact of the national vaccine program.

The seroprevalence of HBsAg was lowest among residents of Penghu, Kinmen, and Lienchiang, whereas HBeAg seroprevalence was highest in these areas. This region also had a relatively lower proportion of mothers with foreign nationality in the birth cohort compared to other regions. The higher HBeAg seroprevalence may be associated with limited access to healthcare in these island areas as well as frequent interaction with people from lower vaccination rates. However, these findings should be interpreted cautiously, given the very small sample sizes for both the numerator and denominator in Penghu, Kinmen, and Lienchiang. Importantly, the results might suggest a higher proportion of active hepatitis B among pregnancies in this region relative to other areas in Taiwan. This observation warrants further research, as it could have implications for other age groups and highlights the need for enhanced monitoring and follow-up of hepatitis B patients in this region.

Preventing MTCT of HBV is a crucial component in the strategy to eliminate viral hepatitis. Notably, the presence of HBeAg signifies high viral replication and infectivity, serving as a strong indicator of active liver disease. Without intervention, infants born to mothers positive for both HBsAg and HBeAg have a > 90% likelihood of contracting HBV at birth, and 80%–90% of these infected infants become chronic HBV carriers [27]. Therefore, the WHO recommends additional preventive measures alongside the HBV vaccine for mothers who are HBeAg-positive or have a high HBV DNA viral load. These measures include administering infant HBV immunoglobulin shortly after birth and providing maternal peripartum antiviral prophylaxis to enhance protection against HBV transmission [28].

In this study, we used logistic regression analysis to examine the effect of birth cohort on HBsAg seroprevalence. While the random forest is highly effective for identifying the strength and direction of associations, especially in capturing complex non-linear relationships, our hypothesis-driven study focused on evaluating the impact of the national vaccination policy on birth cohorts. Logistic regression was chosen for its straightforward interpretability and relevance to our target audience. We recommend that future studies consider advanced statistical approaches, such as random forest, to explore potential non-linear trends in the birth cohort effect.

Our study had several potential limitations. Firstly, data retrieved from health systems might contain misinformation; however, our large sample size should mitigate the impact of this type of information bias. Secondly, data on HBsAg testing was collected retrospectively, and there was no specific HBV vaccination status or dates for vaccination at birth or revaccination for each pregnant woman. Nonetheless, given the high coverage of the nationwide HBV vaccination program, we assume that most of our study subjects born after 1986 received HBV vaccination at birth or in childhood. Also, protective antibody titers may progressively decline after HBV vaccination. However, studies have demonstrated that a large proportion (90%) of individuals with anti-HBs levels below 10 mIU/mL exhibit an anamnestic response following a booster dose, indicating that additional boosters are generally unnecessary for the immunocompetent population [29]. Nonetheless, such studies have not been conducted specifically on pregnant individuals, leaving it unclear whether they might benefit from an additional vaccine dose. Thirdly, we excluded women who underwent prenatal HBV screening at clinics because of the extremely low (< 1%) coverage of HBV testing in the study period (2016–2021). Therefore, selection bias might have occurred, and our results should be interpreted cautiously in the context of pregnant women receiving prenatal services at clinics. Finally, our findings may be inflated due to participants’ high disease awareness, leading them to seek regular check-ups. Conversely, they could also underestimate the situation since foreign women with lower socioeconomic status might primarily attend clinics for prenatal care.

Conclusions

This study highlights the profound impact of national vaccination programs in substantially decreasing HBV prevalence among pregnant women. It underscores the ongoing need for sustained vaccination efforts and the implementation of consistent public health strategies to further alleviate the burden of HBV infection. Beyond vaccination programs, addressing HBV prevention and treatment for foreign immigrants emerges as a crucial aspect that requires attention and resolution.