Age-specific prevalence of hepatitis B virus infection in young pregnant women, Hong Kong Special Administrative Region of China

Abstract Objective To investigate the age-specific prevalence of hepatitis B virus (HBV) infection in young pregnant women in Hong Kong Special Administrative Region (SAR), China, and to determine whether an increase in prevalence occurs during adolescence. Methods HBV prevalence was quantified using data from routine antenatal screening for hepatitis B surface antigen (HBsAg) in 10 808 women aged 25 years or younger born in Hong Kong SAR and managed at a single hospital between 1998 and 2011. The effect on prevalence of maternal age, parity and birth before or after HBV vaccine availability in 1984 was assessed, using Spearman’s correlation and multiple logistic regression analysis. Findings Overall, 7.5% of women were HBsAg-positive. The prevalence ranged from 2.3% to 8.4% in those aged ≤ 16 and 23 years, respectively. Women born in or after 1984 and those younger than 18 years of age were less likely to be HBsAg-positive (odds ratio, OR: 0.679; 95% confidence interval, CI: 0.578–0.797) and (OR: 0.311; 95% CI: 0.160–0.604), respectively. For women born before 1984, there was no association between HBsAg carriage and being younger than 18 years of age (OR: 0.60; 95% CI: 0.262–1.370) Logistic regression analysis showed that the prevalence of HBsAg carriage was influenced more by the woman being 18 years old or older (adjusted OR, aOR: 2.80; 95% CI: 1.46–5.47) than being born before 1984 (aOR: 1.42; 95% CI: 1.21–1.67). Conclusion Immunity to HBV in young pregnant women who had been vaccinated as neonates decreased in late adolescence.


Introduction
It is generally held that vertical transmission is the major route of hepatitis B virus (HBV) infection in regions where the disease is endemic. [1][2][3][4][5] The neonatal immunization programmes using immunoglobulin and hepatitis B vaccination that have been developed to prevent vertical transmission [6][7][8][9][10] have widely been reported to be effective. [11][12][13][14][15][16] Neonatal immunization was introduced to the Hong Kong Special Administrative Region (SAR), China, in 1983 1 and was selectively applied to neonates born to mothers found to be carrying the hepatitis B surface antigen (HBsAg) on routine antenatal screening. 17 In November 1988, neonatal HBV vaccination has become universal irrespective of maternal HBsAg status. 18,19 Since then, vaccination has been readily available from general practitioners and nongovernmental organizations. 17 In Hong Kong SAR, pregnant women receiving antenatal care constitute the only social group that undergoes routine HBsAg screening. We reviewed published reports of the prevalence of maternal HBsAg carriage in the area and found that, in the prevaccination era, it was 6.6% in 1976 20 and 7.4% between 1981 and 1983. 1 In 1996, it was 10.0% and, between 1998 and 2001, it was 8.0%. 21 At our hospital, it was 10.1% in women who gave birth between 1998 and 2008. 22 In our latest survey in 2010, the prevalence was 9.1% overall and 4.8% in women who had undergone HBV vaccination. 23 Since the obstetric population is a low-risk group compared to those with traditional risk factors for acquiring HBV infection in adulthood, such as intravenous drug users and men who have sex with men, pregnant women can be taken as surrogates for the population at large. Consequently, the 9.0-10.0% prevalence of HBsAg carriage observed in these women suggests that Hong Kong SAR remains a high-endemicity area according to the World Health Organization's definition; despite more than two decades of universal neonatal vaccination. 13,[15][16][17] Literature reports suggest that immunity conferred by the vaccine may not always be life-long. One study in Taiwan, China, reported that 1.3% of 2-to 6-year-olds who had undergone complete vaccination were HBsAg-positive, 24 whereas another reported that 1.3-3.5% of immunized infants became HBsAg carriers 15 years later. 25 Furthermore, in Alaska, United States of America, a precipitous decline in antibody titre was observed around the age of 15 years in children vaccinated at a very young age 26 and anamnestic responses to booster vaccinations were absent in half of 15-year-olds vaccinated at birth. 12 Even among fully vaccinated individuals in Taiwan, the seropositivity rate for antibody to HBsAg declined from 100.0% at the age of 2 years to 75.0% at the age of 6 years. 24 In addition, 22.9% of vaccinated children in Israel had undetectable antibody levels, irrespective of gestational age, birth weight or parental origin. 27 Indeed, it has been recommended that a single booster dose should be given 10 years after primary vaccination because protection was estimated to last between 7.5 and 10.5 years. 28 A recent study in Pakistan found that the HBV infection rate rose from 13.4% among individuals aged 11 to 20 years to 34.9% among those aged 21 to 30 years. 29 We observed recently that the prevalence of HBV infection in university students in Hong Kong SAR, increased with age: it was 0.9%, 2.3%, 4.3% and 5.5% in those aged ≤ 18, 19, 20 and ≥ 21 years, respectively. 30 The prevalence also increased with age among women who underwent antenatal screening at our hospital: it was 2.5%, 2.7%, 8.8% and 8.0% in those aged ≤ 16, 17, 18 and 19 years, respectively. 31 These findings suggest that immunity against HBV infection wanes in late adolescence, which could explain the persistently high prevalence of HBsAg carriage we Objective To investigate the age-specific prevalence of hepatitis B virus (HBV) infection in young pregnant women in Hong Kong Special Administrative Region (SAR), China, and to determine whether an increase in prevalence occurs during adolescence. Methods HBV prevalence was quantified using data from routine antenatal screening for hepatitis B surface antigen (HBsAg) in 10 808 women aged 25 years or younger born in Hong Kong SAR and managed at a single hospital between 1998 and 2011. The effect on prevalence of maternal age, parity and birth before or after HBV vaccine availability in 1984 was assessed, using Spearman's correlation and multiple logistic regression analysis. Findings Overall, 7.5% of women were HBsAg-positive. The prevalence ranged from 2.3% to 8.4% in those aged ≤ 16 and 23 years, respectively. Women born in or after 1984 and those younger than 18 years of age were less likely to be HBsAg-positive (odds ratio, OR: 0.679; 95% confidence interval, CI: 0.578-0.797) and (OR: 0.311; 95% CI: 0.160-0.604), respectively. For women born before 1984, there was no association between HBsAg carriage and being younger than 18 years of age (OR: 0.60; 95% CI: 0.262-1.370) Logistic regression analysis showed that the prevalence of HBsAg carriage was influenced more by the woman being 18 years old or older (adjusted OR, aOR: 2.80; 95% CI: 1.46-5.47) than being born before 1984 (aOR: 1.42; 95% CI: 1.21-1.67). Conclusion Immunity to HBV in young pregnant women who had been vaccinated as neonates decreased in late adolescence.
observed in pregnant women. However, the situation in Hong Kong SAR is complicated by the fact that, since the return of sovereignty to China in 1997, there has been a steady influx of immigrants from the mainland, where the prevalence of chronic HBV infection ranged from 4.5% to 17.9% before HBV vaccination was introduced in 1992. 32 Consequently, there was a small but steady addition of infected individuals to the pool of fertile women in Hong Kong SAR. Nevertheless, the extent to which these women contributed to the sustained high prevalence of maternal HBV infection in the region remains unclear. 17,[21][22][23] We hypothesized that the main reason for the current high prevalence of maternal HBsAg carriage was the progressive waning of vaccineconferred immunity against HBV infection in late adolescence. Therefore, we re-examined the age-specific prevalence of positive antenatal HBsAg screening test results in pregnant women who had been vaccinated to determine if and when the age-specific prevalence of HBsAg carriage increases during adolescence.

Methods
Our hospital department is one of the eight public obstetric units in Hong Kong SAR that provide free obstetric care to local residents. We serve a population of 1.2 million, which comprises one seventh of the region's population, and, as the hospital is a tertiary referral centre, we accept referrals from throughout Hong Kong SAR. Our annual delivery rate of 7000 infants is the highest among the eight units. Obstetric management is protocol-driven and includes routine antenatal screening for HBsAg and immunity against rubella, in accordance with official local health policy. The results of investigations and pregnancy outcomes are recorded in a computerized database that covers all public hospitals in Hong Kong SAR. Data entry is performed by trained midwives and obstetricians and is doubledchecked after delivery. To determine whether and when a transition to the current high prevalence of HBV infection occurred during adolescence, we performed a retrospective cohort study of the age-specific prevalence of HBsAg carriage in mothers aged 25 years or younger at delivery who were born in Hong Kong SAR and who were managed between 1998 and 2011. We used data from our obstetric database, which has been previously validated. 22 The study was approved by the Joint CUHK-NTEC Clinical Research Ethics Board.
First, we analysed the age-specific prevalence of HBsAg carriage in the entire group. We investigated the effect on prevalence of the mother's year of birth (i.e. before or in or after 1984, when HBV vaccination became available) and parity (i.e. nulliparity or multiparity) and identified the age at which the transition from a low to a high prevalence occurred. Subsequently, we performed further analyses based on the age of transition identified and, by taking into account the influence of being born before or in or after 1984, we investigated the effect of vaccination on prevalence below and above the age of transition. Statistical analyses were performed using the χ 2 test and odds ratios (ORs) and 95% confidence intervals (CIs) were calculated as appropriate. Correlations between prevalence and age were evaluated using Spearman's correlation coefficient. Multiple logistic regression analysis was used to determine whether the prevalence was significantly influenced by parity, maternal birth before the implementation of HBV immunization or maternal age above or below the age of transition. Calculations were performed using SPSS Statistics 20 (IBM, Armonk, United States of America).

Results
Of the 93 306 women who gave birth between 1998 and 2011 at our hospital, 10 808 (11.6%) were aged 25 years or younger. The overall prevalence of HB-sAg carriage in these young women was 7.5%. As only 129 were aged 16 years or younger, they were grouped together for the analysis. The age-specific prevalence of HBsAg carriage is shown in Fig. 1. There was a significant difference between age groups (P = 0.020) and a positive correlation between age and prevalence (P = 0.006). The age-specific prevalence for women who were born before and in or after 1984 is also shown in Fig. 1 and in Table 1. For those born before 1984, there was no significant difference between age groups (P = 0.558) and no correlation between age and prevalence (P = 0.666). For those born in or after 1984, there was a significant difference between age groups (P = 0.018) and a significant positive correlation with age (P < 0.001).
The most marked increase in prevalence occurred around the age of 18 years in women born both before 1984 and in or after 1984: there was a twofold and a greater than threefold increase between the ages of 17 and 18 years in the two groups, respectively.  Therefore, a cut-off age of 18 years was adopted for the analysis of the effect of vaccine availability on the prevalence of HBsAg carriage ( Table 2). The prevalence was found to be significantly lower in women born in or after 1984 than in those born before 1984: 5.8% versus 8.4%, respectively (P < 0.001). The figure was significantly lower in women born in or after 1984 among both those aged under 18 years (1.2% versus 5.2% for the two birth year groups, respectively; P = 0.025) and those aged 18 years or more (6.2% versus 8.4%, respectively; P < 0.001). However, the difference in prevalence between women aged under 18 years and those aged 18 years or more was significant only for those born in or after 1984 (1.2% versus 6.2%, respectively; P = 0.002). When the effect of parity was analysed, the only significant difference in prevalence found was between nulliparas born in or after 1984 and those born before 1984: 5.6% versus 8.2%, respectively; P < 0.001, Table 3). There was no significant difference in multiparas between those born before or in or after 1984 and no difference between nulliparas and multiparas in either birth year group.
The relative effect on the prevalence of HBsAg carriage of the year of the mother's birth and the mother's age at parturition was evaluated by multiple logistic regression analysis, with parity as a confounder. Being aged 18 years or more had a greater effect on prevalence (adjusted OR, aOR: 2.80; 95% CI: 1.46-5.47) than being born before 1984 (aOR: 1.42; 95% CI: 1.21-1.67). The effect of parity was not significant.

Discussion
The long-term protection provided by a vaccine can be assessed from the anamnestic response to a booster dose, the infection rate in the vaccinated population, in vitro tests of B-and Tcell activity and seroepidemiological studies. The HBV vaccine is thought to induce life-long immunoprotection 11,13,14 and the latest review suggests there is no need for a booster dose in immunologically competent persons. 33 This view is supported by data showing that 97.0% of children demonstrated an anamnestic response to a booster dose 10 years after infancy vaccination, even if there was only a protective antibody concentration in 64.0% of the children. 34 Nevertheless, other studies reported that effective protection lasted only 15 to 20 years. 12,25,26,35,36 Even among fully vaccinated children, the seropositivity rate for antibody to HBsAg has been reported to decline. 24 In one Israeli study, the proportion of children with an undetectable antibody level varied with the interval between vaccination and testing: it was 36.1%, 20.0% and 14.6% in   38 In Taiwan, vaccine failure, which occurred at a rate of 33.3-51.4%, was a major cause of hepatocellular carcinoma: the rate of the disease only declined from 0.54 to 0.20 per 100 000 children aged 6 to 14 years from before to after the introduction of the vaccination programme. Paradoxically, the risk of the disease was higher in HBV carrier children born after the vaccination programme (risk ratio: 2.3 to 4.5). 39 Thus, it is questionable whether the HBV vaccine offers long-term protection. In our study, we observed an agerelated increase in the prevalence of HBV infection in women undergoing routine antenatal screening. Although we could not be absolutely certain all these women had completed a course of HBV vaccination, it is likely they had for the following reasons. In addition to the neonatal vaccination programme in Hong Kong SAR, catch-up vaccination was also given to all school students born before 1983. 40 Moreover, in July 1992 the vaccination programme was extended to cover all children born between January 1986 and November 1988 so that, by the end of the 1990s, all children aged 13 years and younger and residing in Hong Kong SAR, irrespec-tive of place of birth, would have been immunized. 41 Finally, a supplemental HBV vaccination programme was launched in the 1998 to 1999 school year for primary 6 students who had not received or completed the threedose regimen -the aim was to protect all children before they became sexually active. 42 Indeed, our data indicate that the vaccine conferred a protective effect because the prevalence of HBV infection detected by antenatal screening was significantly lower in women born in or after 1984 than in those born before 1984. However, the prevalence of HB-sAg carriage increased with age even in these protected women. In Hong Kong SAR, the prevalence of HBV infection before the introduction of vaccination was 6.6% in 1976 20 and 7.4% in 1981 to 1983. 1 Thus, the 6.9% prevalence of HBsAg carriage we observed in 18-yearold pregnant women born in or after 1984 was comparable to the 1976 figure and the 8.0% prevalence we observed in 22-year-old women exceeded the figure for 1981 to 1983. Moreover, the 8.4% prevalence we observed in 23-year-old women was the same as in the locally born, resident, adult population in 1996. 17 Consequently, our observation of a substantial infection rate in a vaccinated population, taken together with other similar reports in the literature, 25,26,35,36,38,43 provides compelling evidence that the HBV vaccine does not provide long-term protection.
Before mass HBV vaccination was introduced, a study found that the rate of HBV infection increased from 1.7% among children aged 3 to 5 years to 4.5% among teenagers aged 17 to 19 years, 44 which suggests that horizontal transmission is important during the transition to adulthood. In addition to transmission through sexual intercourse and high-risk behaviours, 45 such as the sharing of needles among heroin addicts, horizontal HBV transmission can occur within the family and community in high-endemicity areas because HBV is present in body fluids such as saliva and urine. 43,46 One study found that 17.5% of offspring were HBsAg-positive if either parent was HBsAg-positive compared with 1.5% if both parents were HBsAg-negative 43 and that the presence of an HBV-infected family member was a risk factor at the individual level. This is important in Asia where eating and living with in-laws after marriage is common. It could be that exposure of individuals with declining immunity to infection was one reason for the increase in the prevalence of HBsAg carriage from 1.9% in adolescents to 4.9% in those aged 20 to 30 years reported in the Republic of Korea 43 and for our earlier observation that the prevalence of HBV infection rose rapidly among teenagers in Hong Kong SAR. 30,31 The logical way to reduce the risk of horizontal transmission to individuals with waning immunity is to give a booster dose of HBV vaccine. However, there is no consensus on the necessity of booster doses after neonatal immunization. 12 More than a decade ago it was recommended that 12-year-old children should receive a single booster dose 10 years after primary triple-dose vaccination because protection was estimated to last only 7.5 to 10.5 years. 28 Indeed, one study found that the geometric mean decay in the titre of antibody to HBsAg between the ages of 7 and 16 years in children who did not receive booster doses was 20.0% per year. 47 Moreover, there is growing evidence that the antibody titre declines with age, especially after the age of 15 years. 12,[24][25][26][27]47 The latest guidelines of the Steering Committee for the Prevention and Control of Infectious Diseases in Asia recommend a booster dose of HBV vaccine: (i) 10 to 15 years after primary vaccination in Asian populations where the virus is highly endemic and where it is not feasible to monitor antibody levels; (ii) in immunocompromised patients when the antibody titre falls to below10 mIU/ mL; and (iii) in health-care workers. 48 A booster dose was also suggested for individuals who have a poor response to the vaccine and for high-risk adolescents. 49 Nevertheless, a single booster dose may not be sufficient as it has been reported  25,34 probably because it did not induce an immune response in healthy adolescents who had an undetectable antibody titre. Responses can be impaired by factors such as ethnicity and substance use. 49 There is no consensus on the timing of the first booster dose after neonatal vaccination. Reports of declines in antibody titre and protection 10 to 15 years after childhood vaccination, 12,23,[25][26][27]47 of a precipitous decline in antibody titre at the age of 15 years 26 and of absent anamnestic responses in half of 15-year-old children vaccinated at birth 12,34 all suggest that the immunoprotection induced by the vaccine cannot be guaranteed beyond 16 years of age. In China, the HBsAg seropositivity rate has been reported to increase from 0.3% in individuals aged 15 to 17 years to 1.4% in those aged 18 to 19 years, eventually reaching 3.0% in those aged 24 years. 50 Overall, the evidence -especially our finding that there was a transition to a high prevalence of HBV infection at 18 years of age -supports the recommendation that a booster dose should be given to all adolescents aged between 16 and 17 years. 27,47,48 At present, neither Hong Kong SAR nor any Asian country recommends or gives a routine booster to all adolescents.
Although our study population was limited by the fact that our hospital serves only one seventh of the total population of Hong Kong SAR, the district served comprises mostly new towns that have been developed in the past 20 years to accommodate younger people with new families moving from other districts. Consequently, the young pregnant women in our study are likely to be representative of those who live in other districts and, therefore, of the local female population.
In conclusion, our study found that the prevalence of maternal HBV infection in Hong Kong SAR increased from a low to a high rate around the age of 18 years. Above that age, the prevalence of HBsAg carriage in young mothers who should have been protected by neonatal HBV vaccination returned to that observed in 1976 before the vaccination programme was implemented. Since the pregnant women in our study were not at a high risk of acquiring an HBV infection through horizontal transmission, our findings challenge the view that HBV vaccination confers life-long protection. Future studies should investigate the timing and extent of waning immunity against HBV infection with age, the anamnestic response rate and the most cost-effective way of maintaining immunity in adolescents vaccinated in infancy. ■