Key Findings

Abstract

1. Introduction

HBV is a partially double-stranded DNA virus whose reverse-transcription step contributes to substantial sequence diversity. Genotypes and subgenotypes are geographically structured and are useful for molecular epidemiology, evolutionary analysis, transmission reconstruction, and interpretation of viral diversity. Ethiopia occupies an important position within the Horn of Africa, yet its HBV molecular evidence base remains comparatively small.

A large multicentre Ethiopian study published in 2016 characterized HBsAg-positive blood-donor samples from five geographic regions. Of 391 samples, 371 were successfully analyzed, with genotype A accounting for approximately 78% and genotype D for 22%. The study identified A1 as the genotype-A lineage and substantial diversity within genotype D [4].

Subsequent studies expanded molecular characterization to HIV-positive populations, blood donors, liver-disease cohorts, pregnant women, and occult HBV infection. A 2024 review of HBV molecular epidemiology in Ethiopia reported that only seven primary studies had addressed this field at the time of its search, emphasizing the limited but evolving evidence base [13].

The present analysis was therefore designed to answer a focused question: among successfully molecularly characterized HBV isolates from Ethiopia, what proportions are assigned to the major genotypes and subgenotypes, and how consistent are those distributions across studies?

2. Methods

2.1 Review framework and reporting approach

The work used systematic evidence-synthesis principles informed by PRISMA 2020 and PRISMA-S [1,2]. Because a complete archived database export and exact record-by-record PRISMA identification log were not available for retrospective reconstruction, no fabricated flow counts are reported and the article is titled a systematic evidence synthesis and meta-analysis rather than a completed PRISMA-audited systematic review.

2.2 Research question

Among people with molecularly characterized HBV infection in Ethiopia, what is the distribution of HBV genotypes and subgenotypes, and how does this distribution vary across study populations, geographic settings, infection phenotypes, and genotyping methods?

2.3 Search concepts

Search concepts combined terms for HBV, genotype/subgenotype or molecular epidemiology, sequencing/phylogeny, and Ethiopia. Citation chaining from eligible primary studies and the 2024 Ethiopia molecular-epidemiology review was used to identify linked reports and overlapping cohorts.

(hepatitis B virus OR hepatitis B OR HBV)AND(genotype* OR subgenotype* OR genetic diversity OR molecular epidemiology OR phylogenet* OR sequenc* OR recombinant*)AND(Ethiopia OR Ethiopian)

2.4 Eligibility criteria

2.5 Unit of analysis and denominator

The primary denominator was the number of successfully genotyped HBV isolates, not all screened participants and not all HBsAg-positive persons. This distinction is essential because PCR amplification and sequencing success differed among studies.

2.6 Cohort overlap

Publications derived from the same or overlapping molecular sample sets were not counted as independent prevalence datasets. The later D10 whole-genome report was treated as a characterization of selected Ethiopian strains rather than as an additional independent prevalence study [11]. Earlier overlapping Northwest Ethiopia sequence reports were consolidated into the broader Belyhun 2022 molecular dataset where appropriate [9].

2.7 Statistical analysis

Study-level genotype proportions were calculated using successfully genotyped isolates as denominators. Individual-study 95% confidence intervals were calculated using the Wilson binomial method. Primary pooled estimates for genotypes A and D were generated on the logit scale using random-effects meta-analysis. Between-study variance was estimated with the Paule–Mandel method. Because only five independent datasets contributed to each primary analysis, Hartung–Knapp inference was used for pooled confidence intervals, consistent with the need for caution in few-study random-effects meta-analysis [3].

Heterogeneity was summarized with Cochran’s Q, I², and τ². Approximate 95% prediction intervals were calculated using a t-based small-sample approach and should be regarded as exploratory because only five studies were available. Leave-one-out analyses assessed whether any single study drove the pooled estimate. Meta-regression and formal funnel-plot asymmetry testing were not emphasized because the number of independent studies was too small for reliable interpretation.

3. Evidence Base and Study Architecture

4. Quantitative Meta-analysis

4.1 Genotype A

The proportion of genotype A ranged from 53.8% to 84.8% across the five primary studies. The approximate 95% prediction interval was 29.3%–93.6%, illustrating the large uncertainty expected across future comparable populations. Because only five studies contribute, this prediction interval should be interpreted cautiously.

4.2 Genotype D

The genotype D proportion ranged from 15.2% to 38.3% among the principal sequence-based studies. The approximate 95% prediction interval was 6.8%–62.7%, again demonstrating that the pooled point estimate should not be treated as a universal national proportion.

4.3 Study-level proportions

4.4 Visual distribution across studies

4.5 Leave-one-out sensitivity analysis

When each study was removed in turn, the pooled genotype A estimate remained between 68.3% and 74.0%, while the pooled genotype D estimate remained between 22.8% and 28.5%. No single study reversed the main conclusion that genotype A predominates and genotype D is the second major lineage.

5. Subgenotype Diversity

5.1 Genotype A: strong predominance of A1

A1 is the most consistently reported genotype-A lineage in Ethiopia. Hundie et al. identified A1 within genotype A in their multicentre dataset [4]. Belyhun et al. classified 109 of 110 genotype-A isolates as A1 and one as A9 under the nomenclature used in that study [9]. Dagnew et al. reported all 39 genotype-A isolates as A1 [10].

5.2 Genotype D: greater internal diversity

Genotype D displays substantially greater subgenotypic heterogeneity. In Belyhun et al., the 69 genotype-D isolates were classified as D2 (44), D4 (15), D1 (6), D3 (3), and D10 (1) [9]. Dagnew et al. reported D2 (2), D4 (1), and D10 (4) among seven genotype-D isolates [10]. Earlier Ethiopian work reported D1, D2, D4, historically classified D6 sequences, and a distinct D cluster that was later investigated with whole-genome analysis [4].

The later whole-genome study by Hundie and colleagues characterized selected Ethiopian strains and proposed the designation D10 for a distinct genotype-D lineage [11]. Because those sequences were selected from previously characterized Ethiopian material, that publication is molecularly important but is not counted as a new independent prevalence dataset.

6. Occult HBV: A Distinct Secondary Synthesis

Two Gondar studies selected populations specifically for occult HBV infection (OBI) and therefore were not combined with the main sequence-based genotype pool.

7. Methodologically Distinct 2024 RFLP Evidence

Weldemariam et al. analyzed 7,789 samples, identified 815 HBV-positive specimens, and genotyped 630. The study reported frequent mixed categories, particularly A/E/F (67.3%), using an RFLP-based approach [12].

These categories are not directly interchangeable with sequence-based phylogenetic studies in which an isolate is generally assigned to a mutually interpretable genotype or recombinant lineage. Pooling the RFLP mixed categories directly with the A/D sequence studies would therefore violate the assumption that all studies are estimating the same outcome in the same way.

8. Critical Appraisal and Applicability

A JBI-informed appraisal framework was used conceptually to examine sample frame, recruitment, coverage, molecular measurement, and applicability. No arbitrary numerical quality cutoff was created. The dominant limitation across studies was representativeness, not the mere absence of molecular methods.

8.1 Sequenceability bias

The molecular denominator may select for samples with sufficient HBV DNA for successful amplification and sequencing. Consequently, low-viremia infections can be underrepresented. The correct interpretation is therefore that the pooled estimates describe the distribution among successfully genotyped isolates, not all HBV-infected people in Ethiopia.

9. Discussion

The central finding is robust in direction: genotype A predominates in currently available sequence-characterized Ethiopian HBV datasets, while genotype D is the second major lineage. The random-effects pooled estimates are approximately 71% for genotype A and 26% for genotype D. However, both analyses show substantial heterogeneity, and that heterogeneity is scientifically meaningful.

The included studies sampled different populations—blood donors, pregnant women, people living with HIV, and liver-disease cohorts—and were conducted in different geographic settings. Their molecular denominators also differed. These factors make it unlikely that a single pooled proportion could fully represent HBV genotype distribution throughout Ethiopia.

There is evidence of geographic structure. In the multicentre Hundie dataset, genotype D was relatively more frequent in northern sampling sites than in several central or southern locations [4]. Belyhun et al. also reported a comparatively high genotype-D proportion in Northwest Ethiopia [9]. This pattern is compatible with a north/northwest signal, but the current number of independent regional datasets is too small for a reliable regional meta-regression.

A1 is strikingly consistent across genotype-A datasets. Genotype D is more diverse and includes multiple reported subgenotypes. The Ethiopian D10 lineage is molecularly notable because whole-genome analysis supported a distinct genotype-D cluster [11]. Its epidemiologic prevalence, however, cannot be inferred by double-counting the selected strains used in the characterization paper.

The OBI findings are also important. Genotype D dominated two OBI-enriched cohorts from Gondar [7,8]. This may reflect viral biology, regional ecology, selected population characteristics, or assay-related factors. Larger studies specifically designed to compare OBI and overt HBV within the same regions and sampling frames are needed before a genotype–phenotype association can be claimed.

The 2024 RFLP report further demonstrates why laboratory method matters in molecular epidemiology. High frequencies of mixed A/E/F assignments cannot simply be merged with mutually interpretable sequence-based genotype categories [12]. Future Ethiopian surveillance would benefit from standardized whole-genome or high-quality deep-sequencing approaches capable of distinguishing true mixed infection, recombination, and assay classification effects.

10. Strengths of the Synthesis

11. Limitations

• Only five independent datasets were sufficiently comparable for the primary meta-analysis.
• Most studies used selected populations rather than nationally representative population sampling.
• Several Ethiopian regions remain poorly represented by molecular data.
• Partial S/polymerase sequencing predominates; complete genomes are less common.
• Successful genotyping can select against low-viremia infections.
• Subgenotype nomenclature and classification frameworks have evolved over time.
• Prediction intervals are unstable with only five studies.
• Meta-regression and formal publication-bias tests are not reliable with this small evidence base.
• Exact PRISMA identification and screening counts were not reconstructed; therefore this web publication does not present itself as a completed PRISMA-audited systematic review.

12. Research Priorities for Ethiopia

13. Conclusion

The available sequence-based evidence establishes a clear molecular pattern: HBV genotype A is the predominant genotype among successfully characterized Ethiopian isolates, with a pooled random-effects proportion of 71.1% (95% Hartung–Knapp CI 54.6%–83.4%). Genotype D is the second major lineage, with a pooled proportion of 25.9% (95% CI 15.4%–40.3%).

A1 overwhelmingly dominates the genotype-A evidence base. Genotype D is more heterogeneous and includes D1, D2, D3, D4, and the reported Ethiopian D10 lineage. Rare C, E, and G genotypes have also been detected in selected cohorts.

The high heterogeneity in the main meta-analyses is a warning against interpreting the pooled percentages as fixed national genotype prevalences. Geographic variation, population selection, sequencing success, and laboratory methodology all influence the observed distribution.

14. Publication and Methods Declarations

15. References