Background
Despite the availability of an effective vaccine, infection with hepatitis B virus (HBV) remains a major worldwide health problem: over 2 billion people have been in contact with the virus, and there are 400 million chronic carriers and 1 million deaths per year [
1]. It is hyperendemic in sub-Saharan Africa and Asia [
2‐
5]. HBV is the type species of the
Hepadnaviridae family and is categorized into eight genotypes (A to H) on the basis of a divergence of more than 8% in the entire nucleotide sequence of the viral DNA. Recently, two additional genotypes, I and J, were tentatively proposed [
6,
7].
There are several routes of HBV transmission in sub-Saharan Africa. The highest prevalence is reported in 20–40-year-olds, and horizontal transmission in early life, as a consequence of close family contact, is the most common route of infection [
8,
9]. In the Central African Republic (CAR), E is the prevalent genotype among HBV-infected patients, although genotypes A1, D4 and a genotype E and D recombinant have also been reported [
10].
Most studies of the seroprevalence of HBV in the CAR have been conducted in urban areas, mainly Bangui [
10‐
12], with only a few in rural areas [
13,
14]. In order to obtain a more precise idea of the impact of HBV infection in rural communities, we performed a survey in four prefectures of the country.
Methods
Study population
A cross-sectional study was carried out between November 2007 and April 2008. On the basis of an estimated HBsAg prevalence of 13% and a precision of 4%, we calculated that a minimum of 271 patients were needed. Hence, dried blood spots (DBS) were collected as previously described [
11] from 273 apparently healthy individuals in the prefectures of Ouaka (n = 48), Ouham (n = 75), Nana Mambéré (n = 75) and Lobaye (n = 75). The inhabitants of the four prefectures represent one tenth of the rural population of the CAR. The area selected for sampling in each prefecture was representative of at least half (Lobaye), a third (Nana-Mambéré and Ouaka) and a quarter (Ouham) of the total population. The people living in the study areas were informed of the purpose of the study well before sample collection. Participants were recruited on a voluntary basis. To avoid collecting blood from several members of the same family, donors were selected randomly from among volunteers in each family and from among all those free of symptoms of hepatitis. Most of the participants were illiterate, and several socio-professional groups were represented. Once air-dried, the filter papers were sealed in plastic bags, taken to the laboratory at the Institut Pasteur de Bangui and stored at -20°C with a desiccant until testing.
A questionnaire on socio-demographic characteristics, such as gender, age, place of residence, education, marital status, parity, socioeconomic level and sexual practices, was completed for identification of risk factors for HBV infection. Each participant or his or her parents signed an informed consent form and was informed about the serology results. The study protocol was approved by the Scientific Committee of the Faculty of Health Sciences of the University of Bangui, CAR.
Serological tests
Blood was eluted from DBS as previously described [
11]. DBS were screened with the HBsAg Version 3, anti-HBc (total) and anti-HBc (IgM) antibody kits (Abbott-Murex Biotech Ltd, Dartford, Kent, United Kingdom). Serological tests were performed and interpreted according to the manufacturer’s recommendations.
Detection of HBV DNA and HBV genotyping
DNA was extracted from the DBS with the QIAamp DNA Blood Mini Kit (Qiagen, Venlo, Netherlands) according to the manufacturer’s protocol. The extracted DNA was used for PCR amplification and sequencing as previously described [
15]. Partial sequences covering the
preS and
S gene regions were aligned and compared with SeqScape v2.5 (Applied Biosystems, Nieuwerkerk, Netherlands) and BioEdit version 7.0.9.0 [
16]. Phylogenetic trees were constructed with MEGA4 software [
17] and the neighbour-joining and Kimura 2-parameter methods. Sequences representative of known genotypes and sub-genotypes and the most similar previously published sequences identified by BLAST were downloaded from GenBank and included in the analysis. The nucleotide sequence data reported in this paper are available under GenBank accession numbers JQ972779–JQ972830.
Statistical analysis
Stata 11.0 software was used to determine the prevalence rates of HBV; the Fisher exact test was used to compare variables and the association between HBV positivity and risk factors such as marital status, years since first sexual intercourse, use of condoms, number of sexual partners, socio-professional activity and previous risk behaviour. These risk factors of HBV infection were assessed by estimating odds ratios. Confidence intervals were calculated at 95%. Statistical significance was assessed at p < 0.05.
Discussion
The overall prevalence of HBV detected in the present study was slightly higher than that reported in blood donors in Burkina Faso (14.96%) [
18], lower than those reported in pregnant women in the United Republic of Tanzania (56.2%) [
19] and in Ghana (67.1%) [
20], in a cohort of students in Bangui (42.3%) [
11], in blood donors in Cameroon (86.8%) and Sudan (36%) [
21,
22] and in the staff of a public hospital in Luanda (79.7%) [
23]. The diversity of the observed prevalence is probably due to the diversity of patients and population groups and differences in sampling method, test kit sensitivity and specificity or differences between rural areas and large cities. In large cities, population movement and a permissive lifestyle increase the risk for exposure to this disease.
Dried blood spots are now commonly used for serology and molecular biology testing. The only problem encountered so far is a slightly lower sensitivity, although reliable results are obtained. The technique permits surveys to be conducted in remote areas, as it requires little equipment, the cost of sampling is low, and it is widely accepted as it is almost painless [
11].
A better estimate of the prevalence of HBV infection in rural areas throughout the country would have been obtained if we had surveyed all 16 prefectures; however, this was not possible because of the absence of practicable roads and wide insecurity. Nevertheless, the prevalences of HBV in the four prefectures studied (Lobaye 29%, Nana-Mambéré 28%, Ouaka 29% and Ouham 23%) are similar and lower than the only previous prevalence data published more than two decades ago in the prefecture of Ouham-Pendé (61% in 1984 and 48% in 1988) [
13]. Similarly, the HBsAg prevalence was lower than in Ouham-Pendé (20.9%) [
14] in all prefectures except for Ouaka (19%, Figure
1). The high HBsAg prevalence in Ouaka prefecture was not statistically different from that in the other three prefectures, probably because of the relatively small number of people investigated. Individuals positive for antibodies against HBc but HBsAg negative were found commonly (16.5%), and viral DNA was detected in five. Thus, the number of patients with active HBV infection is underestimated if only HBsAg prevalence is considered. Inability to detect HBsAg may be due to a mutation (G145R for example, which abolishes HBsAg specificity); however, such mutants are unlikely to be selected by vaccination [
24], which in the CAR is given only to young children within the expanded programme of immunization (EPI) . A high prevalence of HBV DNA in the absence of HBsAg was previously described among students in Benin, which was not explained by mutations in the
S gene sequence [
25], and has also been observed previously in the CAR [
10]. A certain characteristic of the population that favours occult infection is a possibility and requires additional investigation [
26]. Occult infections are often overlooked, as these patients are considered uninfected on the basis of HBsAg test results. As a consequence, a diagnosis of chronic hepatitis B should systematically include screening for HBV DNA when antibodies to HBc are present and HBsAg is not.
Analysis of the socio-demographic data did not reveal any significant risk factor for acquiring HBV infection in rural areas of the CAR. Nevertheless, small traders were more often affected than people in other professions (50%); sexual practices (30–40%), tattooing (29%), blood transfusion (13%) and dental surgery (10%) were other possible risk factors. These percentages were much lower than in a cohort of students in Bangui [
11] (sexual practices, 38.2–53.3%; tattooing, 42.4%; blood transfusion, 39.4%; dental surgery; 38%). It is possible that other unknown or unstudied risk factors contribute to the transmission of HBV in rural CAR, and large-scale studies should be performed, including more risk factors.
In contrast to other studies [
4,
21,
23] that showed a certain diversity of HBV genotypes in Central African countries, this study confirms that genotype E is highly predominant, not only in Bangui [
10] but also in rural areas of the CAR, supporting the hypothesis that this country is part of the vast genotype E crescent, spanning Africa from Senegal to Angola [
25]. Characterization of HBV nucleotide sequences from a large sampling in rural CAR should be undertaken to better evaluate the genetic variability of this virus in these areas.
Acknowledgements
The authors wish to thank Professor Elisabeth Heseltine for critical reading of the manuscript and help in revising the written English and Aurélie Sausy for help in performing the molecular biology experiments. We also thank the rural communities of Lobaye, Nana Mambéré, Ouaka and Ouham for their cooperation and are grateful to the health agents at the hospitals of Mbaïki (Lobaye), Bambari (Ouaka), Bouar (Nana Mambéré) and Bossangoa (Ouham) for their help in blood sample collection.
Competing interests
The authors have neither commercial interest in the present study nor any conflict of interest. The Institut Pasteur de Bangui and the Ministry of Foreign Affairs in Luxembourg supported the study financially.
Authors’ contributions
NPK designed the study and directed it, supervised all the field activities, analysed and interpreted the data and wrote the manuscript. UV helped to supervise the field activities, contributed to the acquisition of data by preparing the questionnaire and by participating in the collection of dried blood spots. JMH analysed the molecular data and contributed to writing the manuscript. AB contributed to the acquisition of data by performing ELISA tests. AM prepared the questionnaire and analysed and interpreted the epidemiological data. CPM and ALF contributed to the analysis and interpretation of the data and to writing the manuscript. All the authors approved the final version of the manuscript.