Hepatitis B virus infection among pregnant mothers and children after the introduction of the universal vaccination program in Central Vietnam

A birth cohort study was conducted in Khan Hoa Province, central Vietnam between 2009 and 2012 to determine the seroprevalence of hepatitis B virus (HBV) in pregnant women and their children, and associated risk factors. We enrolled 1987 pregnant women with their babies at the birth phase, and 12.6% (95% confidence interval [CI]: 11.1–14.0) of mothers were hepatitis B surface antigen (HBsAg)+. At 2-year follow-up phase, 1339 (67.4%) children were enrolled of whom 76.6% completed hepatitis B vaccines (HepB) and 1.9% (95% CI: 1.2–2.7) were HBsAg+. When mothers were hepatitis B e antigen (HBeAg)+, 28.3% of children have got infected even with complete HepB. HBV infection in mothers, hepatitis B surface antibody (anti-HBs antibody) below the seroprotective level in children, and mothers with pre-pregnancy low body mass index were associated with HBV infection in children. Meanwhile, HBV infection in children, older maternal age, no or incomplete doses of HepB, and boys were associated with anti-HBs antibody below the seroprotective level in children. Our birth cohort study determined a low rate of congenital HBV infection and associated risk factors in Vietnam, however further studies are needed to advance prevention including anti-viral therapy in pregnant women at high risk.

Of 166 children born to HBV-infected mothers, 122 received complete HepB doses and 44 received incomplete HepB doses. 13 [24.7-81.9]) of children who received complete or incomplete HepB doses got infected. 14.3% (176/1232) of children showed anti-HBs antibody below the protective level despite receiving HepB3. Children born to mothers with higher HBV DNA copies in maternal blood at birth were more likely to get infected with HBV. Children born to mothers with 6-9 Log IU/mL of HBV DNA copy number were more likely to get infected compared to those born to mothers with 4 Log IU/mL defined as a reference, and the higher HBV DNA copy number mothers presented, the higher OR of HBV-infected children ( Table 2). Mean HBV DNA copy number in mothers was statistically higher in HBV-infected children (OR 7.6 [7.1-8.1]) than in non-infected children (OR 4.9 [4.6-5.2]). Among 3 children who received HBIG at birth, 2 children born to HBeAg+mothers had 7-8 Log IU/mL of HBV DNA titer: one with complete HepB doses escaped from infection and the other with incomplete HepB doses got infected. The third child was born to an HBeAg-mother and did not get infected despite incomplete HepB doses.
Factors associated with HBV infection and anti-HBs antibody below the seroprotective level in children. Maternal

Discussion
This hospital-based prospective cohort study demonstrated the seroepidemiological features, risk factors associated with HBV infection, and immunity to it in children, providing implications for the current preventive strategy. This study demonstrated that 12.6% [11.1-14.0] of mothers were HBsAg+, and 42.1% of HBsAg+mothers were HBeAg-. These rates were slightly higher than the previous findings in Vietnam; 9.5% of pregnant women were HBsAg+ 10 , and 16.4-38.2% of adult carriers were HBeAg+ [12][13][14] .
In this study, anti-HBs antibody below the seroprotective level, older ages, and lower educational levels in mothers were associated with current HBV infection of mothers which was in good agreement with the previous studies 12,13 .
Despite the high maternal HBV carrier rate (12.6%), the overall proportion of HBV infection among 2-yearold children was found to be as relatively low as 1.9% [1.2-2.7] with over 90% of children receiving HepB3 and more than three quarters having complete HepB doses. This prevalence corresponds to the one from a previous cross-sectional study reporting HBsAg-positivity rate to be 1.6% among children born in 2007-2008 in Vietnam 16 . Such a level of risk reduction was expected by the previous study projecting 84.8-89.7% of risk reduction in children in reference to 12.5-18.4% in the pre-vaccination era 14 . This was similar to the one in Lao PDR (1.7%) but higher than in Cambodia (0.6%), China (0.3%), and Thailand (0.3%), and lower than in Myanmar (3.8%) in 2012-2017 21,22 . Thus, a successful universal HepB vaccination program could significantly avert perinatal and infantile HBV infection in central Vietnam despite a high-risk condition of vertical and horizontal HBV infection from adult carriers born in the pre-vaccine era.
Our study indicated that 15.2% of children at 2 years of age had anti-HBs antibody below the seroprotective level, which was consistent with other studies suggesting it to be 12.5-20% 23,24 . Such status may mainly include either primary vaccine failure with complete or incomplete HepB, or waning anti-HBs antibody below the protective level by the age. Two-thirds (22/33) of children born to HBsAg+mothers with anti-HBs antibody below the seroprotective level escaped HBV infection possibly due to memory immune to HBV.
HBV infection in mothers, anti-HBs antibody below the seroprotective level in children, and mothers with low BMI in pre-pregnancy were associated with HBsAg-positivity in children. Children born to mothers with higher HBV DNA copy number were more likely to get infected with HBV as previously observed [17][18][19][20] . Obesity may decrease response to HepB 25,26 , no study has found an association of low BMI in prepregnancy with HBV infection in their children. This may be linked to lower socioeconomic disadvantage and health-seeking behavior including immunization. Association between chronic HBV infection and preterm birth were controversial in previous studies 27, 28 and our study revealed no association between them. There is insufficient evidence that CS reduce the risk of mother-to-child transmission of HBV after immunoprophylaxis 29,30 . and our study was in agreement with it. Unlike the previous study 16 , immunization status was not indicated to be statistically associated with HBV infection in children in this study, partly due to the small sample size.
Our study revealed that as much as 28.3% of children born to HBeAg+mothers have got infected even with complete HepB. To prevent mother-to-child transmission of HBV from high-risk mothers further, HBIG at birth would be an option, which is rarely offered only in the private sector mostly in big cities at a high cost. However, some have reported its insufficient efficacy in very high-risk cases. Since only 3 children received it in this study, its efficacy was unable to be evaluated. Instead, anti-viral therapy in pregnant women at high risk has indicated promising evidence on its efficacy and safety 31 as recommended by international liver diseases associations and World Health Organization (WHO) [32][33][34] .
HBV infection in children, older maternal age, no or incomplete doses of HepB, and boys were associated with anti-HBs antibody below the seroprotective level in children in this study. We found that older maternal age was associated with anti-HBs antibody below the seroprotective level in children which was contradictory to a previous finding 20 . High seroprevalence of HBV among women of childbearing age with a slow introduction of the nationwide HepB program may have influenced it. Males responded less to HepB in this study as reported  www.nature.com/scientificreports/ previously [35][36][37] . Moreover, no or incomplete doses of HepB was associated with anti-HBs antibody below the seroprotective level in children, while an association of delayed HepB-BD with immunoprophylactic failure as reported previously 18 was not observed in this study. This study has some limitations. Firstly, approximately 10% of mothers who delivered at either health centers or home and their babies, were not enrolled which could bring some selection bias. Such group of people with restricted medical resources and limited health consciousness are more likely to miss the opportunity of receiving appropriate perinatal and infantile practices including HepB. Thus, the prevalence overall of HBV infection among children could be underestimated, therefore we should be cautious in generalizing the findings. Secondly, due to the logistic and feasibilities issues we were only able to conduct a follow-up survey at 2 years of age, although one to three months after the last dose of HepB would be more appropriate timing for the evaluation. This would have increased the rate of failure to follow-up and the probability of horizontal infection from family or others. We also did not collect the information on possible risk factors linked to horizontal infection such as family history of HBV. These conditions would have made interpretation of the findings difficult. Finally, the immune status induced by HepB might be underestimated since the anti-HBs antibody was measured at 2 years   www.nature.com/scientificreports/ of age when the titers may have waned. Nonetheless, some seroprotective effect was projected among children due to memory immune to HBV even with anti-HBs antibody below the seroprotective level as shown previously.

Conclusions
Our birth cohort study determined a low rate of congenital HBV infection and associated risk factors in Vietnam. Further studies are needed to advance prevention including anti-viral therapy in pregnant women at high risk. HepB vaccination schedule under the national immunization program. The national immunization program for HepB vaccination was revised in 2010 from monovalent HepB-BD followed by 2 times of monovalent HepB at 2 and 4 months of age to monovalent HepB-BD followed by 3 times of pentavalent DPT- www.nature.com/scientificreports/ Hib-HepB at 2, 3, and 4 months of age. The revised program was introduced in June 2010 to KHGH. During the study period, Gene-HBvax (Vabiotech company, Vietnam) was used as a monovalent HepB for HepB-BD, and Quinvaxem (Berna Biotech company, Korea) was used as pentavalent HepB for further HepB doses at KHGH, while vaccines used in health facilities outside the hospital were unknown.

Methods
Subjects and recruitment process. The study targeted residents in 16 out of 27 communes/wards in Nha Trang. Nha Trang is a tourist city, and 16 residential communes/wards were selected as the target area for this study while the remaining 11 communes/wards where big blocks of government administrative buildings, hotels, and restaurants were mostly located were excluded.
In the birth phase, we recruited mothers aged 17 years or older residing in the above areas in Nha Trang who delivered a single child at the hospital on weekdays during this study period without severe maternal complications together with their newborn babies. After we obtained written informed consent from mothers, we collected demographic, clinical, and epidemiological information including self-reported pre-pregnancy body weight through interviews using a structured questionnaire before births as well as from medical charts or maternal health cards. Maternal blood was also collected before deliveries. Clinical status on newborns was examined and recorded after birth. The whole process for collecting information and samples was performed by two trained research nurses under the supervision of a research pediatrician.
In the 2-year follow-up phase, we requested the children recruited in the birth phase to visit their local commune health centers on a designated day within 2 weeks before or after their 2-year birthday. We enrolled those whose caregivers agreed on the condition for further survey including blood sampling. After informed consents were obtained, epidemiological information including the history of immunization was collected through interview using a structured questionnaire and from immunization card. Clinical examination and blood sampling were then conducted and recorded. Health staffs in 16 commune health centers in this study area were carefully trained to perform the prescribed procedures every two weeks, and two trained research staff from Khanh Hoa Provincial Health Service Department supervised them in each center throughout the study period. Vaccination histories were validated through records at commune health centers concurrently.
Sample collection and testing. In the birth phase, plasma samples were separated from maternal blood by centrifugation and kept at − 80 ℃ freezers at the hospital. The samples were then transported to Nagasaki, Japan, where HBsAg and anti-HBs antibody for all, and HBeAg only for HBsAg+cases were examined by chemiluminescent immunoassay (CLIA) (ARCHITECT HBsAg-QT, AUSAB, and HBeAg, Abbott Japan, Tokyo, Japan, respectively). The cut-off values of HBsAg, anti-HBs antibody, and HBeAg were 0.05 International Unit (IU)/mL, 10.0 mIU/mL, and 1.00 S/CO (sample mean chemiluminescent signals [RLUs] /cut-off RLUs), respectively, following the manufacturer's guidelines. HBV DNA titers in mothers were detected by transcriptionmediated amplification (TMA) methods in 3.7-8.7 logarithm genome equivalent (LGE)/ml (DNA probe [FR]-HBV, Fujirebio, Tokyo, Japan).
In the 2-year follow-up phase, blood samples from children were processed similarly as above at Pasteur Institute in Nha Trang and then transported to Nagasaki University, Japan where HBsAg was tested by CLIA (ARCHITECT HBsAg-QT, Abbott Japan, Tokyo, Japan) with 0.05 IU/mL of the cut-off value. Anti-HBs antibody was measured by an assay developed by Green Peptide Co. Ltd., Kurume, Japan using MAGPIX Bead Array System (Luminex, Austin, Texas, USA) with 10 mIU/mL of the cut-off value.
Characteristic definition and categorization. Maternal age was grouped into 17-24, 25-34, or 35-44 years old. Maternal educational level was grouped into low i.e. none or up to junior high school (0-9 years of schooling) and high i.e. high school or higher (10 years of schooling or more). Maternal residential areas were grouped into suburban and urban, defined by a municipal document according to the communes. The numbers of mothers who received ANC 4 times or more, as recommended by WHO 39 , and those less than 4 times were calculated. BMI was defined as the weight in kilograms divided by the square of the height in meters (kg/m 2 ), and underweight was defined as BMI less than 18.50 according to the WHO 40 . Anemia for pregnant women was defined as less than 11 g/dl of haemoglobin 41 , and the preterm is determined as neonates born alive before 37 weeks of gestation according to the WHO 42 .
The protective concentration of anti-HBs antibody was determined as ≥ 10 mIU/mL as indicated by WHO 1 . Anti-HBs antibody levels were categorized into less than 10 mIU/mL (negative or nonimmune to HBV), 10-99 mIU/mL (positive or immune to HBV), or 100 mIU/mL or higher (highly positive or immune to HBV) in this study. HepB-BD was defined as a birth dose of monovalent HepB received within 7 days after birth for this study, and categorized by the timing of receipt into either within 24 h or 2-7 days of life. HepB3 was defined as immunization of a total of 3 or more doses of either monovalent or pentavalent HepB. Immunization status was categorized into complete HepB doses (defined as HepB3 including HepB-BD) or incomplete HepB doses (defined as HepB3 without HepB-BD, or less than 3 doses of HepB with or without HepB-BD). Nonresponse to HepB3 was defined as negative or nonimmune to HBV despite HepB3 given.
Data management and statistical analysis. All the collected information was managed confidentially throughout the process. The data were double-entered and cleaned, and statistical analysis was conducted with STATA 11.1. The study algorithm on enrollment of the subjects was illustrated.
An association of HBV DNA copy number in mothers with HBV infection in their children was analyzed using logistic regression, and tabulated in numbers, proportions, and OR with 95%CI. Here, we used the lowest copy number 4 Log IU/ml as a reference to calculate the OR of HBV infection in subjects with higher copy www.nature.com/scientificreports/ numbers. Mean HBV DNA copy numbers among mothers with HBV-infected and -non-infected children were compared and analyzed using t-test, and ORs with 95%CI were described. Main factors in demographic profiles, clinical-epidemiological factors including seroprevalence, and immunization status were tabulated in numbers and proportions. HBV infection in children was profiled and figured by maternal serological features, and the immunization status of HepB in children, while mean maternal HBV DNA titer was analyzed by HBV infection and immunization status of HepB in children. HBV-infection status with immunization status of HepB in children who received HBIG was described.
Factors associated with HBV infection in mothers, HBV infection in children, or non-response to HepB3 in children were analyzed by univariate analysis respectively using logistic regression. Exposure variables for mothers included HBsAg-positivity, at the seroprotective level of anti-HBs antibody, HBeAg-positivity, age group, ethnicity, educational level, marital status, residential area, previous pregnancy (gravida) and delivery (parity), ANC visit, BMI in prepregnancy, anemia, mode and term of delivery for mothers. Exposure variables for children included HBsAg-positivity, anti-HBs antibody level, sex, low birth weight, small-for-gestational-age, the timing of HepB-BD, immunization status of HepB, and receipt of HBIG at birth. Of these, the main findings were tabulated with OR and their 95%CI. Factors were defined to be associated with HBV infection in children with statistical significance when the 95%CI of OR did not include 1.0. Associated factors were included as exposure variables for further multivariate analysis using logistic regression, being represented with aOR and their 95%CI.
Ethics declarations. This study was approved by the Ethical Committees of Nagasaki University, Japan (# approval no.160908158), and the National Institute of Hygiene and Epidemiology, Vietnam. Written informed consent was obtained from all participants, and all data were kept confidentially and anonymously throughout the process. All methods were carried out in accordance with relevant guidelines and regulations. The informed consent from a parent and/or legal guardian of the participants as minors (mothers who are under 18) are involved in the study.

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.