Hepatitis B vaccination coverage and risk factors associated with incomplete vaccination of children born to hepatitis B surface antigen-positive mothers, Denmark, 2006 to 2010

In Denmark, universal screening of pregnant women for hepatitis B has been in place since November 2005, with the first two years as a trial period with enhanced surveillance. It is unknown what the change to universal screening without enhanced surveillance has meant for vaccination coverage among children born to hepatitis B surface antigen (HBsAg)-positive mothers and what risk factors exist for incomplete vaccination. This retrospective cohort study included 699 children of mothers positive for HBsAg. Information on vaccination and risk factors was collected from central registers. In total, 93% (651/699) of the children were vaccinated within 48 hours of birth, with considerable variation between birthplaces. Only 64% (306/475) of the children had received all four vaccinations through their general practitioner (GP) at the age of two years, and 10% (47/475) of the children had received no hepatitis B vaccinations at all. Enhanced surveillance was correlated positively with coverage of birth vaccination but not with coverage at the GP. No or few prenatal examinations were a risk factor for incomplete vaccination at the GP. Maternity wards and GPs are encouraged to revise their vaccination procedures and routines for pregnant women, mothers with chronic HBV infection and their children.


Introduction
Hepatitis B virus (HBV) infection is a worldwide health problem with more than 350 million people estimated to have chronic liver infections caused by HBV [1]. If hepatitis B surface antigen (HBsAg) is detected in the blood for more than six months, the HBV-infection has become chronic [2]. For infants (up to 1 year old) and children (1-10 years) the two primary sources of HBV infection are perinatal transmission from infected mothers and horizontal transmission from infected household contacts [2]. Mother-to-child transmission of HBV can be effectively (95%) prevented by vaccination [1,2]. The risk of becoming chronically infected with HBV is inversely related to the age of the patient at the time of infection [3]. Chronic infection occurs in ca 90% of infected infants, in 30% of infected children younger than five years and in less than 5% of those infected when they are five years or older [2].
In Denmark, mother-to-child infection is the primary cause (72%) of chronic HBV-infection [4]. Universal screening of pregnant women for hepatitis B has been in place since November 2005 and was made permanent in November 2007 [5,6]. The first two years (1 November 2005 until 31 October 2007) were a trial period with enhanced surveillance. The enhanced surveillance comprised blood banks, maternity wards and general practitioners (GP), for example contacting the maternity wards to secure vaccination of the individual infants and informing the GP about further vaccination of the infants and screening of family members [7,8]. Approximately 180 cases of HBV infection were observed annually between 2006 and 2010, primarily among women from areas highly endemic for hepatitis B [9]. During the trial period, 0.26% of pregnant women were found to be chronically infected with HBV [7].
HBV vaccination is not part of the Danish childhood vaccination programme (CVP) [10,11]. Efforts were instead put into the pregnancy screening programme and into risk group vaccination [12]. This paper presents an evaluation of the pregnancy screening and the subsequent hepatitis B vaccination of the children.
The objective of pregnancy screening is to ensure that all neonates born to HBsAg-positive women are vaccinated against HBV. Furthermore, it is important that GPs refer HBsAg-positive pregnant women to a department for infectious diseases for further information and ongoing care for the newly diagnosed mother including eventually treatment during pregnancy [13]. Only half of the women were referred during the trial period (personal communication: Weiss N, Hvidovre Hospital, Denmark, June 2014, with permission). During the trial period with universal screening, 96% of the children received vaccination at birth [14]. The HBV vaccination coverage of the children born to HBsAg-positive mothers in Denmark during 2008 to 2010 is not known.
According to Danish national guidelines, the vaccination schedule for children born to HBsAg-positive mothers comprises four doses of vaccine, Engerix-B (10 μg): the first dose of vaccine and hepatitis B immunoglobulin (HBIG) at day 0 and three additional doses of vaccine at one, two and 12 months of age [13,15]. All the vaccinations are free of charge [16]. The maternity ward takes care of the first vaccination, including HBIG, and the GPs take care of the rest of the vaccinations.  The aim of this study was to describe the vaccination coverage among children of HBsAg-positive mothers born during the trial period of universal screening with enhanced surveillance (2006)(2007) or born in the following three years with universal screening without enhanced surveillance (2008-2010), to identify risk factors for incomplete vaccination and to discuss possibilities for additional prevention. This article is a proof of concept.

Methods
The study was a retrospective cohort study of pregnant HBsAg-positive women and their live-born children in Denmark from 2006 to 2010. The first two years of the universal screening constituted a trial period with enhanced surveillance.
All Danish residents are registered with a unique identification number (CPR number) [17]. From the national hepatitis B pregnancy screening database (HBpsBA), we know that around 2% of the pregnant women notified with hepatitis B do not have a CPR number. Since they are not listed in the civil register, they were not a part of this study.

Data sources and registers
The national hepatitis B pregnancy-screening database Statens Serum Institute cooperates with the blood banks on monitoring the pregnancy screening in Denmark [15]. The database includes data on all HBsAgpositive women screened since 2005. From this database we extracted all women screened in the defined study period. Data on name, CPR number and date of screening were pulled from the database.

The civil register
This register was established in 1968. Danish residents are registered with a CPR number [17]. From this register, we obtained the CPR numbers of children born in Denmark to mothers known to be infected with HBV and still residing in Denmark. Further, we extracted data on country of origin and immigration/emigration data of the mothers of these women.

The service code register (in Danish: Sygesikringsregisteret)
This is a database containing all services given by primary healthcare practitioners. Each service has a code and each code is registered by CPR number. The database is updated monthly. We extracted service codes for HBV vaccinations of children born to HBsAgpositive mothers (code: 8314-8316), service codes for the prenatal visits (code: 8110-8130) and health examinations of the mothers eight weeks after giving birth (code: 8140).

The national birth register
This register was established in 1973. From this register we identified the hospitals where the HBsAg-positive mothers gave birth.

The maternity wards in Denmark
Data on HBV vaccination and HBIG at birth are registered at the hospitals. All maternity wards were contacted. Data collected were: time of birth, HBV vaccine given (yes/no), HBIG given (yes/no) and if yes for any of these, date and time.

The general practitioners
If data on vaccination were missing in the service code register, the GPs were contacted by phone.  Western Europe (France (n = 1), Spain (n = 2), Sweden (n = 1)) 4 < 1

Study population
We included all HBsAg-positive mothers registered in HBpsBA from 1 November 2005 until 10 January 2011 and the children they gave birth to from 1 May 2006 onwards ( Figure 1).

Definition of 'vaccinated in time'
Scheduled vaccination of children born to HBsAgpositive mothers according to the national vaccination schedule is immediately/within 48 hours after birth, and at the age of one, two and 12 months. We used the following limits for timely vaccination: • First vaccination: immediate vaccination after birth/ within 48 hours • Second vaccination: 5 weeks after birth (nationally scheduled at 4 weeks) • Third vaccination: 10 weeks after birth (nationally scheduled at 8 weeks) • Fourth vaccination: 15 months after birth (nationally scheduled at 12 months)

Risk factors for missing vaccination
We analysed risk factors for incomplete vaccination.
The following risk factors were included: the mothers' country of origin, year of delivery, number of pregnancy examinations, mother's age when giving birth and length of time in Denmark before giving birth. We were able to extract all risk factors from the registers.

Statistical analyses
Kaplan-Meier survival curves were used to estimate the probability for vaccination and how many days after birth the children were vaccinated. The regression model used in this study was logistic regression estimating odds ratios (OR) of included risk factors. The risk factor analyses had three steps: (i) univariate analysis for each risk factor separately, (ii) multivariate analysis with all risk factors included, and (iii) multivariate analysis, final model, with backward elimination with a threshold at 10%. When discussing changes in risk we used a binary regression (Poisson regression without exposure) to estimate risk ratios (RR). We used a significance level of 5%. All data were analysed in STATA version 10.0.

Results
After exclusion of mothers with unknown CPR number, mothers who had given birth outside the defined study period, mothers who had given birth to excluded children and children having left the country, the final cohort comprised 699 children and 594 mothers ( Figure 1).

Characteristics of the mothers
The median age at delivery was 31 years (range: 17-44 years). Of the 594 mothers, 267 (45%) had previously given birth in Denmark. Thirty-five (6%) of the mothers were adopted as children from high-endemic countries for hepatitis B, and 252 mothers (42%) originated from south-east Asia, a high-endemic area ( Table 1).

Vaccination of the children
Of the 699 children, 651 (93%) were vaccinated within 48 hours after birth, 628 (90%) within 24 hours. A total of 305 (44%) of the 699 children had received the second dose of vaccine from their GP at five weeks of age and 304 (43%) had received the third dose 10 weeks after birth. Of the 604 children who had reached the age of 15 months during the study period, 312 (52%) had received three doses of vaccine at the GP. Half of the 604 children (n = 304) had received all four doses (one at birth and three doses at the GP's) ( Table 2). At the time of the study, 95 of the total 699 children had not reached the age for the fourth dose of vaccine or had left the country ( Figure 1) and could therefore not be followed up at age 15 months.
HBV vaccinations were under-reported to the service code register. For 325 of the 699 children (46%), the second dose of HBV vaccine (the first given by the GP) had not been recorded in the register by the GP. However, contact to the GPs revealed that of these 325 children, 251 (77%) had in fact received the second dose. For the third and fourth dose, the corresponding proportions were 64% (208/327) and 36% (166/457), respectively.
Experience from the collection of data in several cases showed that the reason for a disrupted vaccination schedule was that the mother changed GP.   In addition to how many children received the recommended number of doses of vaccine, we also analysed the timeliness of the vaccines given. Figure 2

Risk factors associated with not receiving hepatitis B vaccination within 48 hours after birth
Year of delivery was significantly correlated with risk of not being vaccinated: Giving birth in 2008, just after the enhanced surveillance was terminated, had a 12 times higher risk (relative risk (RR) = 11.70; 95% confidence interval (CI): 2.80-48.80) of not being vaccinated than giving birth in 2007 (Table 3; odds ratio (OR) = 13.06; 95% CI: 2.85-59.90). This increased risk declined in the following years, but was still five times higher (RR = 4.84; 95% CI: 1.14-20.  (Table 3). Neither age of the mother when giving birth nor total time in Denmark before giving birth were significantly correlated with the risk of not being vaccinated. Likewise, the number of pregnancy examinations had no statistically significant influence on the risk of not being vaccinated at birth. Finally, it made no difference in which region of Denmark the birth had taken place, nor was there any difference in the risk of missed vaccination at birth between urban and rural hospitals.
Ten sites of delivery had vaccinated all children, while the remaining 19 sites each counted between one and seven cases of lacking vaccination at birth.
The multivariate models fit well. For the all-factors and final-model analysis in Table 3, the R2 coefficients were, respectively, 0.1349 and 0.1065.

Risk factors associated with less than four hepatitis B vaccinations by the age of 15 months
Newborns who had not received HBV vaccine at birth had a significantly higher risk of incomplete vaccination at 15 months ( Table 4). The number of prenatal examinations was also significantly correlated with the completeness of vaccination, more prenatal examinations being associated with a lower risk of incomplete vaccination. Finally, the mothers' age was significantly correlated with the risk of incomplete vaccination at 15 months: the youngest (17-24 years) and the oldest (35-44 years) age groups of mothers had the highest risk of incomplete vaccination of their children.
Neither the mothers' country of origin nor their total time in Denmark were significantly correlated with the risk of incomplete vaccination. In contrast to vaccination at birth, the year of delivery had no effect on the risk of incomplete vaccination at the GP's (Table 4).
In the multivariate analysis, there was no effect of the enhanced surveillance in 2007 (Table 4). We even observed a slight, but not significant, increasing trend (p = 0.08) over the years in the risk of incomplete vaccination at 15 months of age. Whether or not the mother had had a post-pregnancy examination did not significantly influence the risk of incomplete vaccination at 15 months, although there was a tendency towards a higher risk for those not having received this examination, the RR being 0.87 (95% CI: 0.74-1.02) (p = 0.08).
The multivariate models fit well. For the all-factors and final-model analysis in Table 4, the R2 coefficients were respectively 0.0347 and 0.0237.

Discussion
This retrospective cohort study comprised a total of 699 children born to HBsAg-positive mothers. The participation in the study was 95.2% (four birth records had disappeared or birth site was unknown for 31 children) (Figure 1).

Lessons learned
• Vaccination procedures and routines at the sites of delivery with optimal organisation (best practice) should be a model to implement at all sites.
• The sites of delivery should be attentive to mothers of foreign origin but not forget mothers of Danish origin with chronic HBV infection as well as mothers who were themselves adopted.
• It is important that GPs are informed about vaccinations initiated at the hospital and know the plan for subsequent treatment.
• The GP should pay particular attention to ensure that pregnant women with chronic HBV infection attend all prenatal examinations.
• If it becomes known that a child has not received the vaccination series as recommended, the child should be called in for vaccination.
• If the child changes GP, it is important to communicate the HBV vaccination status to the new GP.
• It is important that the GP employ the special provider number for the vaccine, which forms the basis of any assessment of the vaccination coverage nationally.
The overall vaccination coverage 48 hours after birth was 93% (651/699) ( Table 2). Missing vaccination is considered an adverse event and is subject to internal auditing at the hospital. It is discouraging to find that such a high proportion of the children were not vaccinated according to schedule. Furthermore, 10% (47/475) of the children had not received a single vaccine at the age of two, a dire sign of missed opportunities ( Figure 2 [20]. We expected the Danish birth vaccination percentage to be higher because the standard procedures for this group of women were supposed to be incorporated as routine after the two-year trial period with enhanced surveillance during which the midwives were obliged to fill in questionnaires regarding all children born to HBsAg-positive mothers. In the Netherlands, the Municipal Health Service played a major role in the organisation and follow-up of the children's vaccination status, in contrast to the Danish system. Variation in birth vaccination coverage was observed between the time periods analysed. In the period with enhanced surveillance (1 November 2005 to 31 October 2007) [8], vaccination coverage at the sites of birth was considerably higher than after termination of the monitoring efforts. In 2007, when the routines had become standard procedure at the sites of birth, only 1% did not receive vaccination at birth ( Table 2). Since 2009, the vaccination procedures and coverage have remained at the level they had without enhanced surveillance (94% vaccinated). No effect of the monitoring year was observed for vaccinations at the GPs. We did not find any other studies describing the difference between periods with and without enhanced surveillance.
The salient point is to set a limit for 'vaccinated in time'.
To calculate the vaccination coverage at exactly the date scheduled would give an underestimation of the coverage. Still, the date should reflect whether or not the child has been vaccinated according to the national programme and protected against HBV infection. The time limit for each of the vaccinations (doses two to four) in this study was set based upon the Summary of Product Characteristics (SPC) for Engerix-B [21], minimum/maximum intervals [22] and practicalities around the examination schedules for mother and child. According to the Danish national guidelines, the second dose of vaccine can be postponed until the child is five weeks old [15]. This is also the time for the first (of six) routine scheduled childhood examinations at the GP's. The London study from 2006 looking at risk factors used the limit of 15 months for the fourth dose of vaccine [18]. The same limit was used in our study, an extension of 25% compared with the national schedule that foresees this vaccination at 12 months. We also increased the two other vaccination times with 25% for our study, to respectively five and 10 weeks.
Other studies have focused on setting the limit for 'vaccinated in time' [2,19,[23][24][25]. A study from Thailand found a 3.74 times (95% CI: 0.97-14.39) increased risk of the child being chronically infected with HBV if the interval between the first and second vaccine dose was more than 10 weeks [24]. It has been described that vaccination and HBIG at birth, at one and at six months, compared with at birth, at two and at six months, lead to the same protection from acute and chronic HBV infection in children born to HBsAg-and HBeAg-positive mothers [2]. A study form Switzerland did not find any evidence of age at vaccination (one instead of two months-old) having an influence on prevention of HBV infection [19].
The likelihood of having completed vaccinations at the age of 15 months (but not vaccinations at birth) was positively associated with the mother's number of prenatal examinations, as it was in a study in London from 2006 [18]. Although not significant (Table  3), there seemed to be a slightly lesser risk of not being vaccinated at all, if the mother had no prenatal examinations compared with having one, two or three examinations (Table 3 and 4). This is probably because women with no prenatal examinations are often seen in the maternity ward at birth without being registered before. Thus they have their blood sample taken just before giving birth and/or the newborns are vaccinated without knowledge of the mothers' hepatitis B status.
Ethnicity and total time in the country before giving birth were without significance both in our study and the London study [18]. We were not able to assess the mothers' level of integration into Danish society, nor their command of Danish. In the London study, more than basic command of English was correlated with completed vaccination status. Very few mothers in the study were of Danish origin. The children of these mothers had a significantly higher risk of not being vaccinated than children of mothers from high-prevalence countries. The low prevalence in Denmark could mean that attention was not paid to the small group of HBsAg-positive Danish women. In our study, 6% (35/594) of the mothers were themselves adopted. The hepatitis B status of the adoptees was often unknown by the GP [7]. In Denmark, it is recommended to test adopted children of non-Danish origin for HBV upon arrival in the country [26]. An American study from 2008 found a 4% prevalence for HBV in internationally adopted children [27], which is also high compared to an estimated prevalence of chronic hepatitis B of 0.2-0.3% in the general population in Denmark [14].
Denmark is among the six European countries that have not yet adopted the World Health Organization (WHO) recommendation for inclusion of hepatitis B vaccine in all national vaccination programmes [10][11][12]28]. These six countries (Denmark, Finland, Iceland, Norway, Sweden and the UK) have adopted risk group-targeted hepatitis B vaccination only [29]. In Denmark, even with hepatitis B vaccination included in the CVP, it will still be necessary to maintain the general screening of pregnant women for HBV, with subsequent vaccination of their newborns against HBV. Including hepatitis B vaccination in the CVP would not save the children of hepatitis B-infected mothers from being infected at birth because the first childhood vaccination is not given before the age of three months.
Our study has some limitations. It has not been possible to investigate other potentially relevant risk factors for incomplete vaccination that could help target preventive measures, such as socioeconomic status, written information about hepatitis B given to the mother and whether the hospital record contains information from the GP.
Since no new national initiatives were implemented during the study period, we have no reason to believe that the coverage has improved or the risk factors have changed after the study period. Since January 2014, parts of the enhanced surveillance of the screening have been revived and since 15 May 2014, a national reminder system has been implemented [30]. The reminder concerning the hepatitis B vaccination of children born to mothers infected with HBV will be sent to the GP as soon as possible after birth. Following the results from this study, it is now recommended that GPs test the infants for protective antibody levels after completed vaccination schedule.

Conclusion
In Denmark, timely and complete HBV vaccination of children born to HBsAg-positive mothers need to be optimised. Sites of birth and GPs are encouraged to revise their vaccination procedures and routines for the group of all pregnant women, mothers with chronic HBV infection and their children. Future studies will show if the resumed national monitoring beginning in 2014 will lead to increased vaccination coverage.