HBV immunization and vaccine coverage among hospitalized children in Cameroon, Central African Republic and Senegal: a cross-sectional study

Study population

A cross-sectional study was conducted in five children’s hospitals: one hospital in
Bangui (CAR), two in Yaoundé (Cameroon) and two in Dakar (Senegal). Both the Yaoundé
and Dakar study sites included one pediatric hospital treating infants from families
of impoverished socioeconomic status, and one general hospital serving a relatively
well-off population. The Bangui site was a pediatric hospital serving children from
families of all socioeconomic levels living in the city and its outskirts.

Between April 2009 and May 2010, children aged 3 month to 6 years, hospitalized for
any reason, with a blood sample prescribed during hospitalization, health conditions
allowing an extended blood sample between 2 mL and 5 mL according to the age. Children
were consecutively enrolled after their parents or legal guardians received an information
notice and oral explanation in the local language and provided a written consent.

Ethical approval

This study was approved by the Senegal Health Research National Council, the National
Ethics Committee of Cameroon and the Scientific Committee responsible for validation
protocols and study results in Central African Republic.

Data collection

Data collected were: i) general characteristics (age, sex, weight), ii) clinical features
(reasons for hospitalization, vaccination records on the immunization card), iii)
socio-economic characteristics (place of residence, number of people in the household,
mother’s education (higher level: at least primary education), personal transportation,
electricity, running water, toilets type) and iv) serological data (anti-HBs antibodies,
anti-HBc antibodies, HBsAg, HBeAg) and HBV DNA, when the child was HBsAg-positive.

If the enrolled child’s immunization card was available, vaccination against HBV and
dates of vaccination were recorded. Otherwise, the mother was asked about the child’s
vaccination status.

Complete vaccination was defined as having received all three injections according
to the vaccination card in compliance with the WHO vaccination schedule (6, 10 and
14 weeks of age). Partial vaccination was defined as having received one or two doses
according to the immunization card, regardless of the immunization schedule.

Nutritional status was estimated separately for boys and girls by the Z-score, calculated
on the weight for age, according to WHO standards for children between 3 and 60 months
old, and to CDC standards for older children. Moderate or severe malnutrition was
defined as a Z-score????2 SD 19]–21].

HBV markers

All samples were tested for anti-HBc and quantified for anti-HBs by Enzyme ImmunoAssay
(EIA) (DiaSorin Biomedica, Sallugia, Italy). The correlate of protection for HBV is
an anti-HBs titer ?10 mIU/mL 22], 23]. All children anti-HBs-negative and anti-HBc-positive were tested for HBsAg by automated
EIA (AxSYM, Abbott laboratories, Chicago, USA). All HBsAg-positive children’s mothers
were called by phone so that children could be retested free of charge six months
later. Viral loads were measured by the Cobas AmpliPrep/Cobas TaqMan HBV assay, v2.0
(Roche Diagnostics, Meylan, France) at Saint-Louis Hospital. The limit of detection
was 20 IU /mL. Except for viral load quantification, all laboratory tests were performed
in each country.

Statistical analysis

The children’s characteristics were described as medians and interquartile ranges
(IQR) for continuous variables and percentages for discrete variables.

For univariate and multivariate analysis, quantitative variables were expressed as
dichotomous variables using either the median or a clinically relevant threshold.

Univariate analysis was based on the Fisher’s exact test for discrete variables and
by analysis of variance or the Kruskal-Wallis test for continuous variables. All variables
associated with “having been vaccinated” in univariate analysis (p??0.25) were included in a backward stepwise logistic regression model. A p value
of ?0.05 was considered statistically significant. Adequacy of the model was established
through the Hosmer Lemeshow tests. Interactions between the variables found to be
associated with “having been vaccinated” in the univariate analysis were tested using
likelihood-ratio test. Our data on immunization coverage estimated by serological
markers in children born one year after integration of vaccine into EPI (2005 in Cameroon
and Senegal, and 2008 in CAR) were compared with data reported by WHO on immunization
coverage of surviving infants between 2006 and 2009 24]–26].

Data were analyzed using STATA software version 12.0 (Stata Corporation, College Station,
Texas).

Selected populations for analysis

Using the only publication showing that it is possible to distinguish between the
passive transfer of maternal anti-HBc and HBV exposure in children ?12 months 27], we divided subjects between children younger than 12 months and those older than
12 months.

1) To evaluate the anti-HBV vaccination coverage from serology analysis, anti-HBs?+?and
anti-HBc?+?children??12 months were removed: their vaccination status could not be
determined, since anti-HBs antibodies can be derived from the mother. Anti-HBs?+?and
anti-HBc- children, regardless of age, were considered vaccinated and protected for
an anti-HBs level titer ?10 mIU/mL, assuming that most sub-Saharan African mothers
were unvaccinated [10].

2) To evaluate vaccination coverage by combining serology and vaccination card documentation:
those considered to be vaccinated were children with serological protection (anti-HBs
titer ?10 mIU/mL), as well as children who were unprotected based on serology, but
who had received a complete HBV vaccination according to their immunization cards.

3) To evaluate factors associated with having been vaccinated against HBV (regardless
of serological status, protection or non-protection): an analysis was conducted in
Cameroon and Senegal for children born in 2006 and after, i.e., at least one year
after the integration of the HBV vaccine into the EPIs. CAR children were not included
because the vaccine was integrated later. The variable “vaccinated” implied having
received a complete HBV vaccination according to the immunization card, or if no immunization
card existed, by anti-HBs?+?and anti-HBc- status (with anti-HBs titer ?10mIU/mL).
All anti-HBs?+?and anti-HBc?+?children with no immunization card were removed because
we could not know whether they were vaccinated but not protected, or if they were
not vaccinated at all.

4) HBV current infection was estimated by the number of children with HBsAg-positivity.