Prevalence of HIV and hepatitis B co-infection in Nigeria: a systematic review and meta-analysis

Abstract

Introduction

Hepatitis B virus (HBV) and Human Immunodeficiency virus (HIV) infections are both major public health concerns in sub-Saharan Africa. Co-infection with these virus accelerates liver disease progression, increases hepatotoxicity, undermines treatment effectiveness, and raises mortality risk. We conducted a systematic review and meta-analysis to determine the pooled prevalence of HBV–HIV co-infection in Nigeria.

Methods

We searched databases (PubMed, Scopus, and Embase) from their respective start dates (ranging from 1946 to 1970) through December 31, 2024, to ensure a historical capture of co-infection data in Nigeria. Google scholar was searched to identify relevant grey literature and locally published studies in Nigeria. A random-effects meta-analysis was performed to estimate pooled prevalence and account for between-study variability.

Results

A total of 28 studies were included, resulting in an overall HBV-HIV co-infection prevalence of 7% (95% CI: 5–11%). Heterogeneity was very high (I² = 95.7%), and the prediction interval ranged widely from 1 to 47% resulting in very low certainty of evidence rating. Subgroup analysis showed similar prevalence in adults and children (7% for both). The children’s estimate was homogeneous (I² = 3.7%, 95% CI: 5–8%) with low certainty rating while the adult estimate remained highly heterogeneous (I² = 96.3%, 95% CI: 4–12%;). Regionally, prevalence was higher in the Northern zones 9% (95% CI: 7–12%; I² = 77.2%) than in the Southern zones 6% (95% CI: 3–12%; I² = 96.9%).

Conclusion

This comprehensive systematic review and meta-analysis revealed a significant but highly variable landscape of HIV-HBV co-infection in Nigeria. While stable prevalence was observed in specific subpopulation like children, the high heterogeneity across the broader population suggests that diverse regional and demographic factors might be responsible. Therefore, strengthening integrated HIV–HBV screening and management must prioritize standardized practices, particularly within paediatric care to ensure early intervention and improve surveillance. There is the need for further research to identify the specific drivers of transmission in high-variability areas.

Clinical trial

Clinical trial number not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-026-13320-8.

Introduction

Chronic viral hepatitis, a long-standing inflammatory condition of the liver, remains a silent and neglected public health burden, with viral causes posing a significant threat to population health, particularly in sub-Saharan Africa [1, 2]. Despite its severity and the risk of complications such as hepatocellular carcinoma (HCC), awareness of viral hepatitis infections remains low compared to that of human immunodeficiency virus (HIV) [2, 3]. Among the viral causes of liver disease, hepatitis B and C have the deadliest outcomes [3, 4]. When co-infection with HIV occurs, the risk of mortality and progression to AIDS increases significantly [5]. HIV belongs to the Lentiviridae family [6], whereas HBV is classified under the Hepadnaviridae family [7]. Hepatitis B virus is a major risk factor for hepatocellular carcinoma due to its hepatotropic nature, ability to cause chronic hepatitis, progression to cirrhosis, and direct oncogenic effects, while HIV contributes indirectly to liver disease primarily through immune suppression and co-infection with hepatotropic viruses [8].

HIV and HBV are transmitted through blood and body fluids, including during sexual contact and medical procedures [9]. Although they share similar transmission routes, they target different organs: HIV primarily affects the immune system through the destruction of CD4 + T cells [10], whereas hepatitis B primarily affects the liver, leading to chronic hepatitis, cirrhosis, and HCC [10, 11]. According to a recent meta-analysis, the global prevalence of HBV among PLHIV is estimated at 7.6%, representing approximately 2.7 million cases worldwide [12]. Prevalence varies widely by region and population. Among children, reported HBV–HIV co-infection estimates are 2.6% in the United States and 4.9% in China [13–15]. In sub-Saharan Africa, which remains hyper endemic (≥ 8% prevalence) for HBV according to WHO classification [16, 17], the prevalence of HIV–HBV co-infection ranges from 10% to 20% [13, 18], with overall regional estimates of approximately 15% [19, 20].

Treatment of HIV–HBV co-infection presents significant challenges due to drug resistance arising from overlapping antiviral targets, drug–drug interactions, and the complex interplay between the viruses [20]. Although tenofovir and lamivudine are active against both infections [20], regimen choice is critical. Suboptimal regimens, particularly lamivudine monotherapy, are associated with a high risk of HBV resistance [20]. One of the most clinically important resistance mechanisms involves mutations in the reverse transcriptase region of the HBV polymerase gene, known as the tyrosine–methionine–aspartate–aspartate (YMDD) motif [21]. HIV-associated immune suppression further enhances HBV replication and increases the likelihood of resistance [21, 22]. As a result, high-potency agents such as tenofovir, which has a high genetic barrier to resistance, along with regular virological monitoring and strict adherence, are recommended for co-infected patients [22].

HIV–HBV co-infection is also associated with a greater likelihood of cardiometabolic diseases, driven by chronic immune activation, systemic inflammation, and liver damage from both viruses [23]. These pathophysiologic changes disrupt glucose and lipid metabolism and contribute to metabolic syndrome, while older antiretroviral drugs, especially nucleoside reverse transcriptase inhibitors (NRTIs), have been linked to complications such as insulin resistance and dyslipidemia [23]. Additionally, both HIV and HBV can impair vascular function, contributing to arterial stiffness and increased cardiovascular risk [22].

The financial burden of managing HIV/HBV co-infection is substantial, driven by the costs of diagnostics, antiviral therapy, and treatment of complications such as cirrhosis and HCC [24]. Comprehensive diagnostic testing, including viral load measurement, liver function testing, and resistance assays, adds significantly to costs, while long-term use of tenofovir-based therapies for dual infection contributes further to economic strain [24].

Existing meta-analyses have reported the prevalence of HIV–HBV co-infection in Nigeria; however, these analyses are now outdated and did not formally evaluate the quality of the evidence base. This systematic review provides the most current estimates of HIV–HBV co-infection among adults and children in Nigeria and is the first to apply the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach to evaluate the certainty of evidence. Nigeria is a West African nation with a population exceeding 200 million in 2018 and is divided into six geopolitical zones: the north-central (home to the capital city Abuja), northwest, northeast, and southeast, southwest, and south‒south zones [25]. the North Central (6 states), North East (6 states), North West (7 states), South East (5 states), South South (6 states), and South West (6 states). This regional structure is important for interpreting geographic variation in HBV–HIV co-infection across the country.

It sought to answer the question: What is the prevalence of HIV–HBV co-infection in Nigeria, and how does this prevalence vary across key populations and geographical regions? The resulting high-certainty evidence is critical for informing integrated HIV–HBV screening programs and accelerating progress toward the 95-95-95 targets for HIV and hepatitis B infections outlined by UNAIDS (2020) [24].

Methods

A systematic review and meta-analysis, gathering and analyzing data from previously published studies was conducted. Data was drawn from all six geopolitical zones in Nigeria, reflecting a broad and diverse healthcare landscape. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [26] and the PROSPERO registration number is (CRD420251082795).

Inclusion and exclusion of studies

Studies that reported the prevalence of HBV infection among PLHIV in Nigeria were included. Eligible studies recruited children and adults aged 18 years and older with HIV-HBV co- infection. Co-infection was defined as HBsAg positivity among people living with HIV, diagnosed using Rapid Diagnostic Test or (RDT) or Enzyme-linked Immunosorbent Assay (ELISA) [2–4]. We excluded studies on HBV mono-infection and non-peer-reviewed papers including conference abstracts. Although observational studies, including case-control, cohort, and cross-sectional studies were eligible for inclusion, the search identified a predominance of cross-sectional studies among eligible literature.

Search strategies

A co-author (TJA), systematically searched major bibliographic databases (PubMed, Embase and Scopus) and regional repositories/specialised Repositories (African Journal Online and ScienceDirect) from inception through December 31, 2024. Google scholar was searched to identify relevant grey literature and locally published studies in Nigeria. Searches were tailored to the specific requirements of each database. The search strategy focused on the condition (HBV and HIV), the setting (Nigeria), and relevant study designs. Wildcards, truncation and Controlled vocabularies including Medical Subject Heading (MeSH) for PubMed were utilized and combined using Boolean operators. Example; (hepatitis OR ‘hepatitis B’ OR HBV OR HBsAg OR HIV OR ‘Human Immunodeficiency Virus’ OR ‘HBV-HIV co-infection’) AND (prevalen* OR incidence OR inciden* OR screening OR population OR surveillance OR survey OR ‘observational study’ OR ‘cross-sectional study’ OR cohort OR ‘follow-up study’) AND (‘North Central Nigeria’ OR ‘North East Nigeria’ OR ‘North West Nigeria’ OR ‘South East Nigeria’ OR ‘South South Nigeria’ OR ‘South West Nigeria’ OR Nigeria). See supplementary file for PubMed search strategy. Searches were last updated on December 31, 2024. Additionally, Journal of Hepatitis B and HIV along with the reference lists of identified articles, were reviewed for potentially relevant papers. The study authors were also contacted to address any missing data. All the retrieved articles were systematically organized using the Zotero reference manager.

Screening and study selection

We used Covidence, a software for the management of systematic reviews, to screen articles, remove duplicates and streamline the study selection process. The titles, abstracts and full-text of the identified papers were independently screened by two coauthors (KD and TY) to determine their relevance. Inter-rater reliability for the full-text screening was assessed using Cohen’s Kappa statistic. We resolved any disagreements through mutual consensus, or if necessary, by consulting a third team member (TJA).

Quality assessment

The quality assessment was conducted independently by three co-authors (ANP, NW, and JYB) using a modified 9-point Newcastle–Ottawa Scale [27]. The tool consisted of nine criteria, with one point awarded for each criterion met. These criteria were: (1) the study objective was clearly described; (2) the study design was clearly stated; (3) participants were representative of the source population; (4) participants were accrued during the same time period; (5) the sample size was adequate; (6) missing data were appropriately managed; (7) demographic and clinical characteristics (such as age and gender) were reported; (8) potential confounders, detection methods for HBV and HIV, and sources of bias were described; and (9) study outcomes were clearly defined. Based on these criteria, each study received a score from 0 to 9, with higher scores indicating higher methodological quality. Based on these criteria, each study received a quality score from 0 to 9. Studies were then categorized based on their total quality scores: scores of 0–4 were considered low quality, scores of 5–7 were moderate quality, and scores of 8–9 were categorized as high quality.

Discrepancies were first discussed among the reviewers, and if consensus could not be reached, a fourth senior reviewer (TJA) was consulted to arbitrate. Agreement on the risk of bias judgments across all three independent reviewers was assessed using the Fleiss’ Kappa (appropriate for more than two raters) [28]. The analysis yielded a Kappa value of 0.72 (z = 8.56, p < 0.001), indicating a substantial level of agreement among the reviewers. This shows a strong degree of consistency, suggesting that the reviewers applied the rating criteria in a comparable and reliable manner [29]. Table 1 contains the quality rating of each study, with high quality studies designated as A, moderate quality as B and low quality as C.

Table 1.

Summary of characteristics and quality of included studies (N = 28)

Author (Year) (Ref.)	Geo-Political Zone	Study Design	Testing Method	Sample Size	Total Event (n)	Study Population	Setting	NOS Quality Score
Emuobor et al. 2024 [36]	SW & NC	CS	RDT	537	60	PLHIV	Urban	A
Nwolisa et al. 2013 [37]	SE	CS	RDT	139	8	CLHIV	Urban	A
Oluwaseyitan et al. 2020 [38]	NC	CS	ELISA	281	17	PLHIV	Rural	A
Olanisun et al. 2009 [39]	NC	CS	ELISA	260	30	PLHIV	Urban	A
Cornelius et al. 2019 [40]	NC	CS	RDT	200	7	PLHIV	Rural	A
Adewale et al. 2024 [41]	SW	CS	ELISA	265	2	PLHIV	Urban	A
Ikeako et al. 2014 [42]	SE	CS	ELISA	2500	6	PLHIV	Urban	A
Olusegun et al. 2020 [43]	NC	CS	ELISA	6577	677	PLHIV	Rural & Urban	A
Ifeoluwa et al. 2024 [44]	SW	CS	ELISA	303	12	PLHIV	Urban	C
Taiwo et al. 2012 [45]	SW	CS	ELISA	102	29	PLHIV	Urban	A
Chineze et al. 2023 [46]	SE	NA	ELISA	220	90	PLHIV	Urban	B
Mary et al. 2020 [47]	SW	CS	ELISA	374	20	CLHIV	Urban	A
Olusegun et al. 2021 [48]	SW & NC	CS	ELISA	771	46	PLHIV	Urban	C
Ikpeme et al. 2013 [49]	SS	CS	ELISA	171	10	CLHIV	Urban	B
Uleanya et al. 2016 [50]	SE	CS	ELISA	140	14	CLHIV	Urban	B
Oluwasola et al. 2021 [51]	SW	CS	ELISA	150	13	PLHIV	Urban	A
Oluyinka et al. 2016 [52]	SW	CS	ELISA	182	91	PLHIV	Rural	C
Otegbayo et al. 2008 [53]	SW	CS	ELISA	1779	212	PLHIV	Urban	A
Ejele et al. 2004 [54]	SS	CS	ELISA	342	33	PLHIV	Urban	B
Magaji et al. 2020 [55]	NC	CS	ELISA	3238	408	PLHIV	Urban	A
Aliyu et al. 2015 [56]	NW	CS	ELISA	140	19	PLHIV	Urban	A
Sale et al. 2021 [57]	NE	CS	in vitro	200	22	PLHIV	Urban	A
Kurawa et al. 2024 [58]	NW	CS	ELISA, rapid diagnostic test	100	8	PLHIV	Urban	A
Oboro et al. 2017 [59]	SS	CS	in vitro	295	6	PLHIV	Urban	A
Okonko & Shaibu 2023 [60]	SS	CS	ELISA	104	2	PLHIV	Urban	A
Ojide et al. 2015 [61]	SS	CS	in vitro	342	51	PLHIV	Urban	B
Sadoh et al. 2011 [62]	SS	CS	ELISA	155	12	PLHIV	Urban	A
Uga et al. 2021 [63]	SE	CS	ELISA	384	11	PLHIV	Rural	B

Note: Quality assessment was performed using the Newcastle-Ottawa Scale. Selection (A): Max 4 stars; Comparability (B); Max 2 stars; Outcome/Exposure (C): Max 3 stars. Total score range from 0–9, wheres scores ≥ 7 indicate high quality, 5–6 indicate moderate quality and < 5 indicate low quality. SW= South-West, SS = South-south, SE = South-east, NC= North-central, NW = North-west, NE= North-east. CS = Cross-sectional study

Data extraction

Data extraction form was designed and piloted on a few studies. The form was then adjusted as needed before being used to extract data from the remaining papers. Three co-authors (KED, NW and JYB) independently extracted the following details: first author’s name, prevalence of the study, type of population, method of testing, geo-political zones of Nigeria where the study was performed, study design type, participant demographics (age and sex), method of testing, duration of data collection and total number of subjects. The extracted data were further grouped by region (Northern and Southern Nigeria) and study population (Adults and Children). We contacted study authors where data were missing on an outcome of interest. Disagreements between the data extractors were resolved through discussion with the entire team. This information is presented in Table 1.

Data analysis

Both qualitative and quantitative approaches were considered in this review. For the narrative component, we provided commentary on the included studies, describing their characteristics and assessing their suitability for inclusion in the quantitative synthesis. For the quantitative analysis (meta-analysis), two main statistical models were applied for pooling study results: the fixed-effect and random-effects models [30, 31]. The fixed-effect model assumes that all studies estimate a common underlying effect size and that any observed differences are due to within-study sampling error. In contrast, the random-effects model assumes that the true effect sizes vary across studies due to heterogeneity. The presence of statistical heterogeneity, whether clinical or methodological, does not preclude the conduct of a meta-analysis.

We used the random-effects model with the Restricted Maximum Likelihood (REML) estimator to account for between-study variability, allowing a meaningful pooled estimate to be obtained while acknowledging the underlying diversity in the study populations. To ensure the validity of the statistical model and to stabilize the variance for studies reporting extreme proportions (prevalence near 0), raw estimates were transformed using the Logit transformation (PLOGIT method in R). This approach provided a robust basis for addressing the statistical challenges of proportion data [32] and ensured that differences between studies were appropriately explored and incorporated, yielding a more realistic summary of evidence from diverse studies. This method transforms the bounded proportions (0–1) into a continuous, unbounded scale, thereby stabilizing the variance and ensuring the data conforms to the assumptions of our statistical model [32]. Our assumption for this model was that the true proportions are not identical across all studies and that they vary from study to study due to differences in populations, methods of testing, settings, study periods and the geo-political zones.

The logit transformation mitigates the undue influence of studies with proportions at the extreme ends of the scale, leading to a more robust and statistically reliable pooled estimate [32]. For clinical relevance, the final meta-analysis results, which are initially on the logit scale, are back-transformed to the original proportion scale for a clear and intuitive interpretation of the pooled prevalence or event rate [32, 33]. Most of the proportions in our study were within a moderate range (between 0.2 and 0.8), supporting the appropriateness of the logit transformation [33, 34]. Although no studies in our analysis reported zero events, had this occurred, we would have applied a continuity correction by adding a constant (0.5) to the number of events to allow for their inclusion in the logit transformation and prevent data loss [34]. The pooled logit estimates were then back-transformed to the proportion scale via the inverse logit function and reported with their corresponding 95% confidence intervals [31, 33, 34]. We further assessed the impact these assumptions by conducting the test of heterogeneity and sensitivity analysis where appropriate.

Heterogeneity was evaluated using the I² statistic. We interpreted the I² statistic using the Higgins and Thompson classification, in which values of 25%, 50%, and 75% are denoted as representing low, moderate, and high heterogeneity, respectively [35]. To address the variability between the studies, subgroup analysis was conducted on studies from geographical zones and among specific populations, e.g., adult and children. A leave-one-out sensitivity analysis assessed the stability of the pooled results. Publication bias and small-study effects were evaluated using the DOI plot and LFK index, which provide more objective and sensitive detection of asymmetry compared with traditional funnel plots (102–104). LFK values were interpreted as follows: <1 (no bias), 1–2 (minor asymmetry), and > 2 (major asymmetry). The R statistical software (version 4.5.1), incorporating the metaphor package was used for the main analysis, while Excel was employed to create bar plots.

Results

Study selection process

We identified 12,530 records from multiple databases, including 9,001 from PubMed, 630 from Embase, 910 from Scopus, 1,050 from ScienceDirect, 189 from African Journal Online (AJOL) and 750 from Google Scholar. We also identified an additional 12 studies by screening the bibliographies of eligible studies. Following the initial search, all records were imported into the zotero reference management software (version 7.0) where duplicates were identified using the automated de-duplication algorithm. These results were then manually verified to ensure accuracy before proceeding to the screening stage. After removing duplicates, 3,990 articles remained. Following title and abstract screening, 3,949 articles were excluded because they were irrelevant to the study. This left 43 articles for full-text assessment. See the PRISMA 2020 flow diagram below. Agreement on full-text screening (inclusion/exclusion) decision between the two primary independent reviewers was assessed using Cohen’s Kappa. It yielded a score of 0.72 (p < 0.001), which indicates a statistically significant and substantial level of agreement between the two screeners. For a complete overview of the process, please refer to Fig. 1; the PRISMA flow diagram below.

Fig. 1.

PRISMA flow diagram: Results of literature search and study selection process

Characteristics of included studies

A total of 28 studies with 20,251 participants were included. The studies had a median sample size of 262.5 and recruited adults and children. The adults were a mix of pregnant women, men who have sex with men (MSM), Transgender men (TGM) and farmers. See Fig. 2 below.

Fig. 2.

Composition of included adults participants

The included studies were conducted across six of the country’s geopolitical zones. However, most of the studies lacked sex-disaggregated data, making it impossible to analyze results by sex. Two of these studies covered more than geopolitical zones. See Fig. 3 below.

Fig. 3.

Number of studies by geo-political zone in Nigeria

The full study characteristics can be found in the supplementary material Table 1 (S1). However, a summary of the characteristics of included studies are presented in Table 1 below.

Narrative synthesis

A total of 28 studies were included in this review. However, two studies, Emuobor et al. (2024) and Olusegun et al. (2018), were not included in the regional subgroup analysis because they both included intergeopolitical zone coverage without disaggregating the results. As such, their inclusion could have introduced bias in the subgroup estimates and misrepresented regional prevalence patterns. Nevertheless, these studies contributed to the overall pooled estimate and helped provide a broader national perspective.

Quantitative synthesis

The random-effect meta-analysis revealed a pooled prevalence of 7% (95% CI: 0.05–0.11%). However, heterogeneity was very high (I² = 95.7%, p < 0.0001) as shown in Fig. 4. The 95% Prediction Interval (PI) ranged from 0.01 to 0.47. This wide PI suggest that the true prevalence in a new, unstudied population is highly likely to fall within this broad range.

Fig. 4.

Random effect meta-analysis of prevalence of HBV‒HIV co-infection in Nigeria

Heterogeneity across studies was very high, as indicated by the I² statistic (I = 95.7%; p < 0.0001). The high I² value of 95.7% indicates considerable variability among the studies, suggesting that the prevalence estimates may not be consistent across different populations and settings. Consequently, a Galbraith plot was constructed to visualize and identify studies contributing most significantly to the observed variability. The plot revealed several studies positioned outside the 95% confidence limits, suggesting potential outliers that may have substantially influenced the overall heterogeneity (Fig. 5).

Fig. 5.

Galbraith Plot of study contributions to overall heterogeneity

Meta-regression

A multiple meta-regression model was conducted to examine whether study setting and diagnostic testing methods accounted for the high level of heterogeneity (105). The model showed substantial residual heterogeneity (τ² = 1.69; I² = 98.2%), with an R² of 0%, indicating that neither moderator explained any of the between-study variation. Residual heterogeneity remained highly significant (QE(21) = 542.52, p < 0.001), and the omnibus test of moderator was non-significant (QM(6) = 0.25, p = 0.9997) (106). None of the individual moderator coefficients reached statistical significance (all p > 0.77), suggesting no meaningful differences by study setting or diagnostic assay.

Sensitivity analysis

Influence analysis including Baujat plot, Leave-One-Out analysis and influence diagnostics were conducted to address the heterogeneity observed in the primary analysis (I² = 95.7%). The Leave-One-Out Analysis demonstrated the stability of the pooled effect as the removal of any single study did not result in a significant shift of the overall estimate. See supplementary material 2 (S1 for Baujat plot and influence diagnostics), and supplementary material 3 (S3) for the leave-one –out analysis. The diagnostics identified Ikeako et al. 2014 as the most influential study on the models parameters. See supplementary material 4 (S4) for the serial Forest plots. To quantify the influence of these and other outliers, a series of sensitivity analysis were performed, and the results are shown in Table 2.

Table 2.

Sensitivity analysis of pooled estimate and heterogeneity

Analysis	Studies (k)	Pooled Estimate (Logit/Effect size)	95% CI	t2	I2
Primary Analysis	28	7%	5–11	1.3232	95.7%
		Sensitivity Analysis
(Excluding study 7)	27	9%	6–12	0.8607	95.1%
Excluding study 7 and study 17	26	8%	6–11	0.6279	92.7%
Excluding Studies 6, 11 and 17	25	7%	5–9	0.8685	90.8%
Excluding studies 6, 10, 11, 17 and 24	23	8%	6–9	0.2400	83.5%

Publication bias

Publication bias was assessed using the Doi Plot and its associated statistical measure, the Luis Furuya-Kanamori (LFK) Index. See Fig. 6 below for the plot and its index.

Fig. 6.

DOI Plot with the LFK Index

The calculated LFK index was 0.07, which falls within the range indicative of minor or nor asymmetry. Hence, there was no evidence of significant publication bias in this meta-analysis.

Subgroup analyses

Stratified analyses were conducted by population and geopolitical region. Adults (PLHIV) and Children (CLHIV)) living with HIV had a prevalence of HBV‒HIV co-infection of 7% (95% CI: 4‒12; I² = 96.3%; p < 0.0001) and 7% (95% CI: 5‒9; I² = 3.7%; p < 0.0001), respectively. The test for subgroup differences between the PLHIV and CLHIV populations was not statistically significant (Q = 0.21, df = 1, p = 0.6479).

See Fig. 7 for subgroup analysis by population.

Fig. 7.

Random effect meta-analysis of the prevalence of HIV-HBV co-infection in PLHIV and CLHIV in Nigeria

Regional subgroup analysis

To investigate the influence of geography on the observed heterogeneity, we visually examined the HIV-HBV co-infection prevalence rate using a Bubble plot. The Bubble plot showed that studies from the Northern region exhibited both higher mean prevalence and a significantly wider range of prevalence compared to studies from the Southern region. The heightened prevalence may reflect historical and socio-economic factors affecting vaccination rates and healthcare accessibility. Hence, geographical location contributed to the structural heterogeneity in the primary studies. See Fig. 8 below.

Fig. 8.

Bubble plot showing prevalence rate of HBV-HIV co-infection by Geographical location

The subgroup analysis revealed a greater pooled prevalence in the Northern (N) part of the country 9%; (95% CI: 7–12; I² = 77.2%; p < 0.0001) compared to the Southern (S) Nigeria 6%; 95% CI: 3–12; I² = 96.9%; p < 0.0001). There was no statistically significant difference between the two regions; (Q = 1.18, df = 1, p = 0.2766) in spite of the pattern seen in the Bubble plot. See Fig. 9 below.

Fig. 9.

Random effect meta-analysis of HIV-HBV co-infection in Southern and Northern region of Nigeria

Assessment of certainty

The certainty of evidence for each subgroup was rated according to the five GRADE domains: risk of bias, inconsistency, indirectness, imprecision and publication bias [64, 65]. All bodies of evidence originated from observational studies, given them an initial certainty rating of Low.

Children subgroup

The certainty of evidence for the prevalence in children was rated low. While the results were highly consistent (I² = 3.7%), the rating was downgraded one level due to a Moderate Risk of Bias (40% of contributing studies were rated as moderate quality or less on the NOS).

Adult subgroup

The certainty of evidence for the prevalence in adults was rated as Very Low. This rating was the result of two major concerns: it was downgraded two levels for Severe Risk of Bias (over 30% of studies were rated as moderate or low quality on the NOS), and a further one level for Serious Inconsistency, as the results remained highly heterogeneous (I² = 96.3%).

The Key Findings and final GRADE ratings are presented in Table 3.

Table 3.

Summary of Findings (SoF) - Prevalence of HBV–HIV Co-infection in Nigeria

Outcome (Study Population)	Number of studies/no. of participants	Pooled Prevalence estimate/CI	Certainty of Evidence	Reasons for Certainty Rating
Prevalence of HIV-HBV co-infection in Children	5/979	7%(95% CI: 5–8)	Low	Downgraded once for RoB (Moderate RoB in Primary studies). No downgrade for Inconsistency
Prevalence of HIV-HBV co-infection in Adults	23/19,272	7% (95% CI: 4–12)	Very low	Downgraded 2 levels for RoB in primary studies, and one level for serious inconsistency

Discussion

This review shows that HIV–HBV co-infection remains a significant public health issue in Nigeria. Although the pooled estimate suggests a moderate prevalence, the very wide prediction interval indicates substantial variability across studies. Such variability underscores that the true prevalence likely differs markedly across populations and settings. In a densely populated country like Nigeria, even a 7% co-infection rate translates into a considerable absolute burden and heightened demand on healthcare resources.

The pooled HBV–HIV co-infection prevalence of 7.0% across adults and children was lower than estimates reported in other African systematic reviews, including Cameroon (0.8%) [66], Ethiopia (10.21%) [67] and Ghana (13.6%) [68], underscoring regional heterogeneity in HBV–HIV epidemiology across Africa. Similarly, another African systematic review and meta-analysis reported an overall occult hepatitis B prevalence of 11.2% among HIV-positive individuals, with the highest regional prevalence in southern Africa (26.5%), followed by northern (11%), eastern (9.1%), and western Africa (8.5%) [69].

In contrast, our estimate closely mirrors findings from the 2018 Nigerian HIV/AIDS Indicator and Impact Survey, which reported an HBV prevalence of 8.1% among adults aged 15–64 years [47, 70], suggesting consistency with nationally representative Nigerian data. Compared with other world regions, our pooled prevalence was lower than estimates from China (13.7%) [71] and the Western Pacific Region (27%) [72], but similar to those reported in Latin America and the Caribbean (7.0%) [73], and the global pooled prevalence of 7.6% (98). Taken together, these comparisons indicate that Nigeria’s HBV–HIV co-infection burden falls within the broader international range, although differences across settings likely reflect variations in background HBV endemicity, vaccination coverage, study populations, and diagnostic approaches. The slightly lower prevalence observed in our review may also be explained by the inclusion of both adults and children, who differ in susceptibility, exposure pathways, and behavioural risk profiles.

These disparities likely reflect differences in HBV endemicity, healthcare access, vaccination coverage, study design, and timing of data collection [74]. While our findings align with estimates from Latin America and the Caribbean, the lower prevalence in Nigeria compared to Ethiopia and Ghana suggests regional variations in vaccination and healthcare access, which should be explored further.

A previous Nigerian review by Owolabi et al. (2014) [75] reported a higher pooled prevalence of 15%. This difference may relate to the earlier review’s broader inclusion of retrospective studies, all from urban settings, where higher rates are commonly recorded due to improved diagnostic access and greater population mobility. Our review applied stricter methodological criteria and incorporated updated evidence, yielding a more balanced and contemporary national estimate. These findings reinforce the need for sustained public health attention and integrated screening and prevention strategies [25].

Marked geographic disparities were observed, with the North showing a higher pooled prevalence than the South. This pattern, illustrated by the bubble plot (Fig. 8), may reflect higher HBV endemicity, weaker vaccination uptake, and disparities in healthcare infrastructure across regions [76, 77]. Sociocultural behaviors, limited integration of HIV–HBV services, and differences in study populations and diagnostic methods may further explain these regional variations [78]. Collectively, these findings point to a moderate but substantial national burden of co-infection, with implications for accelerated disease progression.

Heterogeneity was high in the overall and adult analyses, indicating that the pooled prevalence should be interpreted as a summary estimate rather than a precise national rate. Despite exploratory and sensitivity analyses, residual heterogeneity persisted. Literature shows that HIV markedly alters the natural history of HBV infection, accelerating progression to cirrhosis and hepatocellular carcinoma [79, 80]. This necessitates specialized care, including careful selection of antiretroviral therapy (ART), as some agents act on both viruses and may increase hepatotoxicity [15, 81, 82]. A Nigerian study by Idoko et al. [83] found that elevated baseline Alanine Transferase (ALT) levels predicted hepatotoxicity after ART initiation.

Beyond liver complications, co-infected individuals face increased cardiometabolic risks due to chronic inflammation, metabolic dysregulation, and long-term ART exposure [84, 85]. Drug–drug interactions, such as those involving tenofovir-based regimens, further complicate management [86]. A Nigerian study by Muhammed et al. reported high rates of hypertension, obesity, and metabolic syndrome in patients on ART [87]. These clinical challenges highlight the need for integrated care and robust preventive strategies.

From a public health standpoint, the persistent burden of HBV–HIV co-infection underscores the urgency for Nigeria to accelerate its current strategies to meet global elimination targets. The WHO Global Health Sector Strategy (2022–2030) calls for integrated management of HIV and viral hepatitis, and failure to address HBV co-infection undermines progress toward UNAIDS 95–95–95 goals [24]. This review supports two critical shifts in Nigeria’s HIV care framework; mandatory, integrated HBV screening at HIV diagnosis and routinely during care and scaling up HBV vaccination for all HBsAg-negative individuals living with HIV. Implementing these strategies is essential to mitigate liver disease progression, reduce mortality, and advance toward global elimination goals.

The economic implications are also significant. Managing chronic HBV costs approximately ₦564,959 (US$376) annually, and costs rise further with dual infection due to prolonged treatment and monitoring needs [88]. For many Nigerians, these costs are prohibitive, and current funding mechanisms are inadequate. Thus, primary prevention, including vaccination and public education, is the most cost-effective strategy. The moderate prevalence documented here reinforces the need for routine HBV screening in HIV care and widespread vaccination campaigns, especially in high-risk groups. Achieving these goals will require investments in human and material resources amid chronic health system underfunding and low doctor-to-patient ratios [89].

Among children aged 2 months to 17 years, the pooled HBV–HIV co-infection prevalence was 7%, with very low heterogeneity, indicating a robust and reliable estimate. This prevalence was lower than reported in Benin (10%) but higher than in Ivory Coast (5%) and Togo (1%) [90]. the same systematic review also reported a slightly higher overall HBV prevalence of 9% among West African children [16]. Comparatively, lower rates observed in the United States (2.6%) and China (4.9%) [13–15] likely reflect stronger HBV vaccination programs and earlier integration of HBV screening into paediatric HIV care. The relatively higher prevalence in Nigeria may be explained by maternal HBV infection, limited vaccine coverage, and gaps in early childhood HBV screening programs. The referenced studies [44, 60] were cross-sectional and context-specific, whereas our meta-analysis synthesizes multiple data sources, offering a more comprehensive overview.

Transmission patterns also differ by age. Children often acquire HBV vertically or through early horizontal transmission, with a higher likelihood of chronic infection due to immune immaturity [91, 92]. Among adults, HIV transmission predominantly occurs through heterosexual contact [93], whereas HBV acquisition often occurs earlier in life, resulting in sequential infections. This highlights the importance of integrating HBV testing and vaccination across the HIV care continuum and strengthening early-life screening efforts.

Our regional subgroup analysis yielded prevalence estimates consistent with recent studies reporting slightly lower values in the North (8.0%) and South (3.96%) [44, 58]. Differences likely reflect variations in study design, diagnostic approach, and population characteristics. Given that HIV increases susceptibility to chronic HBV, vaccination or booster doses should be offered to all HBV-seronegative individuals living with HIV. Integrated vaccination and screening within HIV care is essential for reducing co-infection rates and advancing national viral hepatitis elimination goals [23, 24].

The persistent prevalence of HIV–HBV co-infection in Nigeria indicates gaps in the implementation of existing HIV and hepatitis policies, particularly in routine HBV screening and consistent use of HBV-active antiretroviral therapy [77, 78]. Strengthening integrated HIV–HBV services, expanding hepatitis B vaccination among high-risk adults, and improving co-infection surveillance are essential to support Nigeria’s HIV control and hepatitis elimination goals [25]. Furthermore, the success of these mitigation efforts is contingent upon robust health education and awareness programs. There is currently a significant knowledge gap regarding the risk of HBV- HIV co-infection among both diagnosed patients and those yet to be screened. Public health campaigns must emphasise that HBV is asymptomatic, so that routine screening is a standard part of HIV care to ensure early detection for undiagnosed individuals. For those already infected, health education should focus on the benefit of HBV-active antiretroviral therapy and the importance of lifelong adherence to prevent severe liver complications.

Strengths

This review provides the most up-to-date evidence on HIV–HBV co-infection in Nigeria using a rigorous and transparent methodology. Comprehensive analyses, including subgroup analysis, sensitivity testing, meta-regression, and DOi/LFK index assessment enhanced the credibility of findings. Inclusion of both adult and pediatric populations allowed meaningful age-specific insights. The GRADE approach strengthened confidence appraisal, identifying the pediatric estimate as particularly robust due to low heterogeneity. Collectively, these strengths make the findings highly relevant for policy and program planning.

Limitations

The primary limitation of this review is the substantial heterogeneity observed, particularly in the overall and adult subgroup analyses. While high heterogeneity is common in prevalence meta-analyses, it nonetheless limits the generalizability of pooled estimates [94, 95]. Sensitivity analyses and meta-regression did not identify sources of the observed variability, suggesting that unmeasured study-level or population-level factors may be contributing. Consequently, pooled estimates should be interpreted with caution.

In addition, most included studies were assessed as having a high risk of bias, which reduced the certainty of evidence. Using GRADE, we rated the overall and adult population estimates as very low certainty, whereas the estimate for children was more robust. Finally, the predominance of cross-sectional study designs limits the ability to infer temporal relationships between HBV and HIV co-infection. Longitudinal studies are therefore needed to better elucidate these dynamics.

Differences between manuscript and protocol

This systematic review and meta-analysis was prospectively registered in PROSPERO with the title “Prevalence and burden of HBV–HIV co-infection in Nigeria.” During peer review, it was noted that our analysis, which synthesized prevalence estimates, did not include metrics required to quantify disease burden (e.g., incidence, disability-adjusted life years, mortality, or economic costs). We therefore revised the manuscript title and text to remove the term “burden” and to more accurately reflect the scope of the study as a meta-analysis of prevalence.

This modification does not affect the objectives, eligibility criteria, search strategy, included studies, statistical analyses, or conclusions of the review conducted in Nigeria. The change was made solely to improve methodological precision and transparency in reporting. The protocol record on PROSPERO remains unchanged because registry entries cannot always be edited after registration; however, the difference between the registered protocol and the final manuscript is explicitly reported here in accordance with PRISMA recommendations.

Conclusions and policy implications

This systematic review and meta-analysis provide updated evidence on the burden of HBV–HIV co-infection in Nigeria. Although the pooled estimate reflects a moderate prevalence, the burden among children is particularly noteworthy and underscores a sustained public health concern. Co-infected individuals face heightened risks of treatment complications, accelerated liver disease, and poorer clinical outcomes. While the pooled prevalence should be interpreted as a summary estimate, given the very low certainty of evidence and substantial heterogeneity. These findings highlight the need for integrated HBV screening within existing HIV care frameworks in Nigeria, particularly targeting high-risk populations to meet national and global elimination targets.

Strengthening prevention, surveillance, and integrated care is therefore essential. Alignment with the WHO Global Health Sector Strategy on Viral Hepatitis (2022–2030) will be critical to advancing progress toward viral hepatitis elimination and improving outcomes for people living with HIV. Sustained, coordinated national action remains central to mitigating the long-term clinical and economic consequences of this dual infection. In light of the evidence synthesized, routine hepatitis B screening should be integrated into existing HIV testing and treatment platforms, including antenatal care and general adult HIV services, to facilitate earlier diagnosis and timely clinical management. Incorporating hepatitis B vaccination into major HIV care programs is an effective and cost-efficient strategy, as demonstrated in other settings, and should be prioritized for high-risk groups such as children, newly diagnosed HIV-positive individuals, and their close contacts. To achieve Nigeria’s viral hepatitis elimination goals, clinical intervention must be coupled with robust health education initiatives. Education represent a highly cost-effective strategy with significant long-term payoffs for public health. Addressing the critical lack of awareness among both infected individuals and those yet to be diagnosed, targeted education can promote screening, enhance treatment adherence and support prevention.

Moreover, expanding HIV surveillance systems to include hepatitis B markers would generate more accurate, nationally representative data on co-infection trends and improve policy decision-making and resource allocation. Future research should address the methodological limitations highlighted in this review by adopting standardized diagnostic criteria, enrolling larger and more diverse populations, and using more rigorous study designs. Such improvements are essential to enhance the quality of evidence and to better guide national strategies for managing HBV–HIV co-infection in Nigeria.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

HBV

Human B Virus

HIV

Human Immunodeficiency Virus

WHO

World Health Organisation

UNAIDS

Joint United Nations Programme on HIV/AIDS

ART

Antiretroviral Therapy

ALT

Alanine Aminotransferase

AIDS

Acquired Immune Deficiency Syndrome

PLHIV

People Living with HIV

HCC

Hepatocellular Carcinoma

CD4

Cluster of Differentiation 4

YMDD

Tyrosine–methionine–aspartate–aspartate

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

NRTIs

Nucleoside Reverse Transcriptase Inhibitors

RDT

Rapid Diagnostic Test

ELISA

Enzyme-linked Immunosorbent Assay

REML

Restricted Maximum Likelihood

Author contributions

JYB conceptualized the study, conducted the initial literature search, and drafted the manuscript. TJA critically reviewed the manuscript, carried out a comprehensive literature search, performed all statistical analysis, and provided guidance on all aspects of the study. Data extraction and quality assessment were conducted by JYB, ANP, KED and NW. KED reviewed the final draft to enhance its overall quality. All authors (JYB, TJA, KED, ANP and NW) read and approved the final version of the manuscript for publication.

Funding

This research did not receive funding from individuals or any organization.

Data availability

The dataset and R script used to perform the meta-analysis are publicly available and can be assessed online in the Zenodo repository. The materials are accessible via the following Digital Object Identifiers (DOIs): Dataset: [10.5281/zenodo.17788688], R Codes: [10.5281/zenodo.17790151].

Declarations

Ethical approval and consent to participate

Not applicable as this study relied on publicly available data.

Consent for publication

All authors read and approved this manuscript for publication.

Competing interests

The authors declare no competing interests.