Prevalence of Hepatitis B Virus Infection in Tanzania: A Systematic Review and Meta-Analysis

Methods We systematically searched the PubMed, Web of Science, African Journals Online, Embase, Cochrane Library, and Google Scholar databases for studies conducted up to March 1, 2023, that estimated the prevalence of HBV in Tanzania based on HBV surface antigen measurements. The DerSimonian–Laird random effects model was used to estimate the overall prevalence of HBV with 95% confidence intervals (CIs). Potential sources of heterogeneity were also investigated. Results Thirty-one studies with a total sample size of 37,988 were included in the meta-analysis. The overall average HBV prevalence estimate in Tanzania was 6.91% (95% CI = 5.18–8.86%). Subgroup analysis revealed the highest prevalence in the northern zone (9.32%, 95% CI; 2.24–20.36%), among the blood donors (18.72%, 95% CI: 17.43–20.05%) and among the community volunteers (8.76%, 95% CI: 4.55–14.15%). The lowest prevalence was observed in the lake zone at 4.66% (95% CI: 3.49–5.99) and in pregnant women at 4.72% (95% CI: 3.42–6.21). The overall between-study variability showed significant heterogeneity (I2 = 97.41%, P < 0.001). Conclusions Our results showed that Tanzania is a country with moderately high HBV endemicity, with large interregional differences and significantly high numbers of HBV infections within the community. This underscores the need for immediate development of targeted prevention strategies and further epidemiological studies to better understand the pattern of the disease.


Introduction
Hepatitis B virus (HBV) infection poses a signifcant global public health burden, with approximately 296 million people living with chronic hepatitis B (CHB) worldwide [1].HBV infection can lead to serious liver diseases, such as cirrhosis and liver cancer.In sub-Saharan Africa, including Tanzania, the prevalence of HBV infection is particularly high.
Available subpopulation studies in diferent parts of Tanzania have shown the prevalence of HBV to be 1.2-11.2%[2].
Despite the signifcant global burden associated with HBV infection, recent estimates indicate that only approximately 10% of people living with this disease have been formally diagnosed, and only 13% of those diagnosed are receiving treatment [3].Tis global estimate coincides with the local data from Tanzania, where less than 10% of the population was screened for HBV, with minimal detection and treatment rates [4].Tis concerning gap in diagnosis and subsequent treatment exposes a substantial number of individuals to an increased lifetime risk of liver-related complications, such as cirrhosis, hepatocellular carcinoma (HCC), and events of end-stage liver disease and death.Indeed, more than 66% of cases in the country with HCC were correlated with HBV infection [5].
Globally, there have been several eforts to address this issue, with the goal of eliminating HBV as a public health threat by 2030 [6].Prevention through vaccination and antiviral treatment is a key strategy for achieving this goal.However, according to the recent report from the World Health Organization (WHO) [7], this target is likely to be missed, particularly in low-and middle-income countries, mainly due to a lack of comprehensive viral hepatitis prevention and control programs and poor surveillance systems for the disease.According to this report, there is still a high number of new infections and deaths; 1.5 million people were newly infected with HBV in 2019, with 1.1 million deaths.
In response to the global HBV elimination goal, Tanzania launched a national strategic plan for viral hepatitis in 2018 [4].However, the efective policy implementation is still limited, as vaccination rates for high-risk groups have remained below the desired level, 2% in pregnant women [8], 74% in healthcare workers (HCWs) [9], and 0% in people living with HIV (PLWHIV) [2].Te WHO [6] recommends the HBV vaccination for all individuals at high risk of HBV infection.Additionally, the number of eligible patients receiving antiviral treatment for HBV is limited in the country [10].Te inconsistency of the data, coupled with insufcient surveillance systems in the country, may be the major factor hindering efective policy implementation, subsequent preventive strategies, and resource allocation.To address this issue, a systematic review and meta-analysis of available studies in Tanzania were conducted to estimate the prevalence of CHB in the country, describing the disease pattern.

Search Strategy.
Tis study was registered with the International Prospective Register of Systematic Reviews (PROSPERO) with registration number CRD42023472128 and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [11].A systematic literature search was conducted in the following electronic databases: PubMed, Web of Science, African Journals Online, Embase, Cochrane Library, and Google Scholar, for studies up to March 2023.For PubMed, the following MeSH terms and texts were used: "Hepatitis B" .Te latter is a list of all administrative regions in Tanzania.Two independent authors (SBK and VDK) screened the titles and abstracts of the manuscripts to determine their relevance, and the full texts of the selected studies were subsequently reviewed.Similar search strategies were used for other databases that included terms such as hepatitis B virus, viral infections, prevalence, Tanzania, and prevalence of hepatitis in the various target groups.

Inclusion Criteria.
Te inclusion criteria were studies conducted in Tanzania, published in English before March 2023, used the hepatitis B surface antigen (HBsAg) test for diagnosing HBV infection, and included participants of any age.

Exclusion Criteria.
Case reports, reviews, preprints, and studies with insufcient or inaccessible data were excluded.

Data Collection and Management.
Mendeley software was used as the reference manager to remove duplicates of the studies that were identifed from the electronic databases and to generate bibliographies.Te titles and abstracts that were produced from electronic databases were independently screened by two authors as per eligibility criteria.Te full texts of the qualifed studies were then reviewed, and the data were extracted and entered into a Microsoft Excel-designed extraction form using the following variables: publication year, study design, study location (regional/zone), population (general, PLWHIV, pregnant women, HCW, people who inject drugs (PWID), and children), study year, sample size, prevalence of HBV, and diagnosis method.A third reviewer was involved in double-checking the correctness of the data entry and resolving disagreements.
2.4.Quality Assessment.Quality assessment was done using the Newcastle-Ottawa Quality Assessment Scale [12].All the parameters of this scale (standardized methods for confrming diagnosis, large sample size, multicenter study, appropriate statistical methods that report results, accounting for confounders, clear methodology of selection of participants, and representativeness of the population) were considered in our meta-analysis.Te quality assessment was carried out by two independent reviewers, and disagreements were resolved by a third reviewer.

2
Journal of Tropical Medicine

Statistical Analysis.
Quantitative data on the overall number of subjects with HBV were extracted, as were the data in the subgroups.Te HBV prevalence across studies was calculated, and due to heterogeneity, the data are presented as ranges of values with 95% confdence interval (CI).Furthermore, weighted means were calculated considering the diferent study sizes.
For the meta-analysis, contingency tables were created in an Excel spreadsheet to compare the prevalence of HBV in diferent study populations.Given the inherent variability among observational studies, we used the random efects model of DerSimonian and Laird to estimate the HBV prevalence in the included studies, as used elsewhere [13].Heterogeneity across studies was assessed using the heterogeneity index (I 2 ); I 2 > 70% suggested high heterogeneity, 50-69 indicated substantial heterogeneity, and <49 indicated low heterogeneity.

Study Identifcation.
Te fowchart for the selection of studies is presented in Figure 1.A total of 120 studies were found in various electronic databases.After removing duplicates, 66 articles remained.After screening the titles and abstracts, 6 more studies were removed and 48 studies were chosen for full manuscript reading.From those, 17 studies were excluded due to missing prevalence (4), conducted outside Tanzania (5), and 8 being reviews leaving 31 studies for the fnal review.Te risk of bias assessment according to the Newcastle-Ottawa Quality Assessment Scale is shown in Supplementary Table 1.

Study Characteristics.
We included all studies that met the eligibility criteria and were conducted before 01 March 2023 in our review.Table 1 shows the characteristics of the studies that were eligible for our meta-analysis.A total of 31 studies were analyzed, with a combined sample size of 37,988.Te duration of the studies ranged from two months to nine years, with the majority being cross-sectional studies (26 out of 31; 83.9%).Approximately one-third of the studies (11 out of 31; 35.5%) were conducted in the eastern zone of the country, specifcally in the Dar es Salaam and Morogoro regions.Te Zanzibar, southern, and central zones had minimal representation, with only one study (3.2%) each.Among the reviewed studies, Hawkins et al. [31] had the largest sample size (17,539), while Machange et al. [23] had the smallest sample size (68).Te majority of the studies (17 out of 31; 54.8%) were conducted between 2013 and 2023, while only 5 out of 31 (16.1%) were conducted between 1991 and 2001.Rapid diagnostic tests (RDTs) were the most commonly used method for diagnosing HBV in 17 (54.8%)studies.

Subgroup Analysis of HBV Prevalence. As shown in
Table 2, our meta-analysis divided the studies into groups based on geographical location, year of study, studied population (PLWHIV, blood donors, HCW, pregnant women, and PWID), and HBV diagnosis method.When considering the geographical location of the country, the highest prevalence of HBV was found in the northern zone, with 9.32% (95% CI; 2.24-20.36%)and a total of 3779 participants.Te eastern zone had the next highest prevalence of 7.93% (95% CI; 5.57-10.66%)with a sample size of 25,487.Te lowest prevalence of 4.66% (3.49-5.99)was observed in the lake zone.Te southern, central, and Zanzibar zones were each represented by one study (Figure 3).

Discussion
Efective policy implementation requires data on the magnitude of disease.Tis meta-analysis review reports the prevalence of HBV in diferent populations in Tanzania from 1994 to 2023.Our fndings showed a pooled prevalence of 6.91%, with variations observed among diferent geographical zones, study years, diagnostic methods, and population subgroups.Tis prevalence is similar to that reported in Ethiopia (6.0%) [13] and the average prevalence in four other East African countries (6.03%).Tanzania had a lower prevalence of 5.16% compared to Kenya and Uganda, which had higher prevalence of 8.54% and 8.45%, respectively [44].Similarly, a neighboring country, Malawi, had a higher prevalence of 8.1% [45].Tese results suggest that Tanzania has intermediate-high endemicity of HBV, while most neighboring countries have high endemicity according to the WHO classifcation [46].Te lower prevalence in Tanzania may be attributed to the government's eforts to improve preventive measures, such as early infant HBV vaccination.Tanzania has higher cumulative vaccination coverage rates (89.6%)compared to Uganda (77.6%) and Kenya (86.7%) [47].Other factors, including HBV genotypes and high-risk behaviors, may also contribute to these diferences.Nonetheless, Tanzania being Journal of Tropical Medicine intermediate-high HBV endemicity, more aggressive public health measures are needed in the country to control HBV and align with global targets for elimination by 2030.
Our research found signifcant diferences in the occurrence of HBV in diferent areas of the country.Studies in the northern and eastern regions reported higher rates than the national average, while lower rates were reported in the lake and southern highlands regions.Tese fndings suggest that there may be variations in the factors that contribute to HBV infection, such as prenatal screening, vaccination rates, community awareness, and the persistence of the infection.However, further investigation is necessary to understand the underlying reasons for this pattern.Similar disparities in HBV prevalence have been reported in other countries, including Iran [48], China [49], and Kenya [50], independent of factors like the studied population or diagnostic methods used.As for the southern, Zanzibar, and central regions, it is challenging to determine the overall prevalence of HBV due to the limited number of studies conducted in each area, with diferent populations and timeframes.In the southern region, Shedura et al. [39] analyzed the prevalence of HBV among pregnant women attending antenatal clinics in 2022.In Zanzibar [26], febrile outpatients were analyzed in 2007 while the analyzed population in the central region included all outpatients attending districts and regional hospitals from 1991 to 1992 [14].
Our study also assessed the prevalence of HBV infection in diferent populations, and one signifcant fnding was the higher prevalence of HBV in the general population (blood donors and community volunteers) compared to high-risk groups such as PLWHIV, HCW, PWID, and pregnant women.Tis contrasts with fndings from other studies in intermediate-and high-endemic areas.For example, in Bangladesh, the national average prevalence of HBV in the general population is 4%, lower than that in PWID (7.5%) and HCW (7.3%) [51].Similarly, in China, a recent metaanalysis reported a national HBV prevalence of 3.8% in the general population, with higher rates in high-risk populations such as PLWHIV (10.7%) and PWID (15.0%) [52].Another study in Sierra Leone found a higher prevalence of 15.9% among PLWHIV compared to the national average of 13.0%, while pregnant women and HCW had lower percentages of 9.7% and 11.9%, respectively [53].One possible explanation for our fndings is that the high-risk populations consistently have access to healthcare facilities, making them more likely to receive health education and vaccination against HBV infection compared to the general population.Previous studies in Tanzania have shown that more than two-thirds of HCW were vaccinated against HBV infection, and general knowledge about HBV was good [54,55].Tese fndings were backed up by another study by Ndunguru et al. [56] that further revealed that knowledge and HBV vaccine uptake among the HCW were signifcantly higher in the urban areas compared to rural areas.However, even though the prevalence in high-risk populations is lower than the community, it is still higher than the national average as found in our study.Terefore, these fndings highlight the importance of public health measures and call for simultaneous eforts to strengthen HBV elimination programs both in the community and healthcare settings.
In terms of the diagnosis methods, the prevalence rate of HBV detected using RDT was found to be lower than that of studies using ELISA.Tis may be due to the lower sensitivity    [57,58].Te performance of RDT can also be infuenced by factors such low levels of HBsAg and HBV deoxyribonucleic acid (DNA), HBV genotypes, specifc brand of HBsAg tests used, and the HIV status of the individual [59].Due to these shortcomings, the WHO recommends using RDTonly in the settings where laboratory testing is limited and/or in populations where access to rapid testing would facilitate linkage to care and treatment [60].In this meta-analysis, the majority (54.8%) of published studies, including those involving PLWHIV, used RDT as a diagnostic method, which may have led to an underestimation of the true prevalence of HBV in our study.

4.1.
Strengths.Tis study employed an extensive search approach across signifcant data sources and included a large number of studies and various subgroup populations from most geographical zones of the country.Consequently, our  fndings provide an accurate representation of the current HBV situation in Tanzania.

Limitations.
One signifcant drawback of this metaanalysis was the uneven representation of certain subgroups such as children and various geographical zones.Only one study [30] focused on the population of infants who are particularly susceptible to HBV infection through maternal transmission.Similarly, only one study was available for some geographical zones such as southern [39], central [14], and Zanzibar [26].Terefore, it is important to be cautious when interpreting the overall fndings of our study.

Conclusion
To the best of our understanding, this study is the frst to provide a meta-analysis on the prevalence of HBV in Tanzania.Te current literature suggests that Tanzania is a country with an intermediate-high endemicity of HBV, but with signifcant variations between regions.Additionally, this study confrms a considerable number of HBV infections within the community.Tese results highlight the immediate need for targeted prevention strategies, such as awareness programs, universal immunization, and relevant policies.Moreover, further research should be conducted to better understand the underlying factors infuencing the observed infection patterns.

Figure 1 :
Figure 1: PRISMA fow diagram for identifcation and selection of articles for inclusion in the review.

Figure 2 :
Figure 2: Forest plot of the studies on HBV prevalence in Tanzania published from 1994 to 2023.

Table 1 :
Characteristics of the included studies in the systematic review and meta-analysis for the prevalence of hepatitis B in Tanzania from 1994 to 2023.

Table 2 :
Subgroup analysis of the estimated prevalence of HBV in Tanzania in the studies published from 1994 to 2023.