This systematic review and meta-analysis were conducted to synthesize published data on the prevalence and antimicrobial resistance of ocular infections in Africa. The study protocol was registered prospectively on the International Prospective Register of Systematic Reviews (PROSPERO) under the registration number: CRD420251127675. The review is reported in accordance with the PRISMA 2020 (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines to ensure transparency and methodological rigor [10].

To capture contemporary microbial profiles and recent trends in antimicrobial susceptibility among ocular isolates, this review considered data published over the last quarter-century. Consequently, studies published between January 2000 and November 2025 were eligible for inclusion.

To ensure comprehensive literature coverage, searches were conducted across multiple electronic databases: PubMed/MEDLINE, Google Scholar, and African Journals Online (AJOL). Additional searches targeted relevant scientific society websites in microbiology and ophthalmology. Furthermore, institutional reports from the World Health Organization (WHO), the Global Action Fund for Fungal Infections (GAFFI), and the Centers for Disease Control and Prevention (CDC) were consulted.

The search strategy employed a combination of Medical Subject Headings (MeSH) terms and keywords, utilizing Boolean operators (AND, OR). For PubMed, the following representative strings were used:

((((((Eye Infections [MeSH]) OR (Conjunctivitis [MeSH])) OR (Keratitis [MeSH])) AND (Microbial Sensitivity Tests [MeSH])) OR (Drug Resistance, Microbial [MeSH])) AND Africa [MeSH]).((((Eye Infections [MeSH]) AND (Diagnosis, Microbial [MeSH])) OR (Anti-Bacterial Agents/Pharmacology [MeSH])) AND Africa [MeSH]).

For AJOL and Google Scholar, broader terms such as “eye infections AND microbiological analysis AND Africa” were applied to capture regional publications.

The studies were included if they met the following criteria: 1) conducted in an African setting, 2) patients with clinically manifest ocular infections characterized by inflammatory signs, purulent discharge, or symptomatic distress, and 3) reported primary data on microbial prevalence, profile, or antibiotic susceptibility.

Exclusion criteria involved: 1) non-primary research (systematic reviews, editorials, conference abstracts), 2) primarily measuring bacterial colonization of the ocular surface (e.g., routine preoperative screenings in asymptomatic surgical patients), 3) studies on non-ocular specimens, 4) data published before 2000, and 5) systemic infections with only secondary ocular involvement.

Reference management and initial de-duplication were performed using Zotero. Records were then imported into the Rayyan platform for blinded screening. Two independent reviewers (M.G.S. and G.S.M.) assessed the methodological quality of the included studies using the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Prevalence Studies. Discrepancies were resolved through discussion or by a third reviewer (M.L.). Potentially relevant articles underwent full-text review. Finally, the methodological quality of included studies was assessed using the Joanna Briggs Institute (JBI) critical appraisal tool for analytical cross-sectional studies [11].

A standardized form was used to extract publication year, country, type of ocular infection, sample source, sample size, number of positive cultures, and antimicrobial resistance rates. We specifically recorded whether resistance was defined according to CLSI or EUCAST guidelines to ensure comparability.

Meta-analysis was performed using RStudio (version 4.5.2) with the meta package (version 7.0 or higher). Pooled prevalence and resistance rates were estimated using a random-effects model (if I² > 50%) or a fixed-effects model (if I² ≤ 50%). Effect sizes are presented with their 95% confidence intervals (CI). Subgroup analyses were conducted by African region (East, West, North, South, Central) and by bacterial species to address the geographical variations mentioned in the introduction. The primary outcome was the bacterial culture positivity rate among clinically suspected cases. To ensure rigor, denominators were standardized by analysis: 1) total samples for overall positivity; 2) identified isolates for pathogen distribution; and 3) specific strains tested for resistance and multidrug-resistance (MDR) rates.

The literature search initially identified 1978 studies. Duplicates were removed using the Rayyan tool. After this step, 1947 unique studies were retained for screening. Following the title and abstract screening, twenty-three (n = 23) studies were deemed fully eligible. Finally, fifteen (n = 15) quality studies were retained. The PRISMA flow diagram (Figure 1) documents the entire process.

This systematic review included 15 cross-sectional studies [12]-[26] published between 2003 and 2024. Ethiopia led research production with 9 publications, followed by Nigeria (n = 3), while Egypt, Eritrea, and Tanzania each had one study. Diagnostic methodology relied primarily on standard microbiological culture. Antibiotic susceptibility was tested primarily by disk diffusion. The Clinical and Laboratory Standards Institute (CLSI) guidelines were the most frequently reported. The main characteristics of the studies are detailed in Table 1.

Figure 1. PRISMA flow diagram of study selection.

Table 1. Characteristics of eligible research papers.

The methodological quality of the included studies was assessed using the Joanna Briggs Institute (JBI) critical appraisal checklist for analytical cross-sectional studies. The results are presented in Table 2. Overall, all 15 included studies demonstrated moderate to high methodological quality, with scores ranging from 7 to 9 out of a possible 9.

Abbreviations: JBI: Joanna Briggs Institute; Q1 - Q9: quality assessment criteria.

Q1: Was the sampling frame appropriate to address the target population?

Q2: Were study participants recruited in an appropriate way?

Q3: Was the sample size adequate? Q4: Were the study subjects and the setting described in detail? Q5: Was the data analysis conducted with sufficient coverage of the identified sample? Q6: Were valid methods used for the identification of the condition? (Identification of bacterial/fungal pathogens) Q7: Was the condition measured in a standard, reliable way for all participants?

Q8: Was appropriate statistical analysis used? Q9: Was the response rate adequate, and if not, was the low response rate managed appropriately? 1 = Yes, 0 = No.

Table 2. Quality assessment of included studies using the JBI checklist.

The pooled analysis of 15 cross-sectional studies comprising 5004 samples estimated an overall ocular infection prevalence of 65% (95% CI: 58% - 71%), with substantial heterogeneity (I² = 97%; p < 0.0001) justifying the use of a random-effects model. Individual study prevalences ranged widely from 33% (95% CI: 30% - 36%) in the study by Barakat et al. [16] to 94% (95% CI: 91% - 96%) in that by Okesola and Salako [26]. The 95% prediction interval (34% - 87%) confirms the considerable dispersion of potential effects (Figure 2).

The pooled prevalence was 61% (95% CI: 56% - 66%) in East Africa [12][14][15][17]-[19][21]-[25] compared to 74% (95% CI: 42% - 92%) in West Africa [13][16][19][20], with corresponding heterogeneity of 86% and 99.2%, respectively. Despite these disparities, the test for subgroup differences was not significant (p = 0.40). The very high overall heterogeneity (I² = 97%) and the prediction interval confirm the substantial dispersion of potential effects (Figure 3).

Figure 2. Forest plot of the pooled prevalence of ocular infections with 95% confidence intervals.

Figure 3. Forest plot of the pooled prevalence in East and West Africa.

The stratified analysis by sample size shows pooled prevalences of 72% (95% CI: 62% - 81%) for small studies (<250), 61% (95% CI: 55% - 67%) [13][17][25] for medium-sized studies (250 - 350) [12][14][18]-[24], and 68% (95% CI: 23% - 94%) for large studies (>350) [15][16][26], with very high heterogeneity across all subgroups (I² = 86.4% to 99.2%). The test for subgroup differences was not significant (χ² = 3.70; p = 0.155), indicating that sample size does not explain the observed variability. Overall heterogeneity remains very high (I² = 97%), confirming the substantial dispersion of results (Figure 4).

Figure 4. Forest plot of the pooled prevalence stratified by study size.

The stratified analysis by type of infection showed a prevalence of 62% (95% CI: 55% - 68%) for external eye infections (13 studies, n = 3831) [12][15][18]-[24][26], compared to 50% (95% CI: 20% - 80%) for keratitis (2 studies, n = 1173) [14][16]. The test for subgroup differences is significant (p = 0.0025), indicating that infection type explains part of the observed variability (Figure 5).

Analysis of the microbiological profile of ocular isolates (n = 3047) revealed a marked predominance of Gram-positive cocci (n = 2122, 69.64%), mainly Staphylococcusaureus (n = 1207, 39.6%) and coagulase-negative staphylococci (n = 566, 18.57%). Gram-negative bacilli constituted the second major pathogenic group (n = 597, 19.59%), dominated by Escherichia coli (n = 134, 4.4%) and Pseudomonas aeruginosa (n = 129, 4.2%), whereas Gram-negative cocci were rare (n = 48, 1.6%). The fungal component (n = 277, 9.09%) was characterized by the predominance of the genus Aspergillus, notably Aspergillus flavus (n = 93, 3.05%) (Table 3).

Analysis of trends over the 2003-2024 period reveals a significant recomposition of the microbiological landscape of ocular infections. This evolution is characterized by a marked decline in Staphylococcus aureus, with its prevalence decreasing

Figure 5. Forest plot of the pooled prevalence stratified by type of infection.

Table 3. Microbial profile of pathogens isolated from ocular infections.

from 45% to 27%, paralleled by a dramatic increase in coagulase-negative staphylococci (CoNS), whose isolation rate increased from 0.0% to 24.5%, as well as a notable increase in Klebsiella pneumoniae. Furthermore, the appearance of fungal pathogens, with the emergence of the genus Aspergillus during the 2019-2021 period, further complicates this evolving picture (Figure 6).

Analysis of the heatmap reveals distinct resistance profiles according to species. Staphylococcus aureus presents a very high level of resistance to β-lactams (approximately 75%) and moderate resistance to macrolides (less than 50%). Pseudomonas aeruginosa displays variable resistance, ranging from 30% to cephalosporins to 100% to macrolides and 75% to β-lactams. Escherichia coli develops a profile comparable to P. aeruginosa, except for cephalosporins, where resistance is very low. Resistance remains moderate against aminoglycosides and fluoroquinolones, with Streptococcus escaping the latter family (Figure 7).

Figure 6. Evolution of the frequency of isolated germs over the years.

Figure 7. Antibiotic susceptibility testing results: percentages of resistant strains isolated from ocular infections, based on multiple studies in Africa.

Among the fifteen studies included in this analysis, eight (comprising 1324 bacterial isolates) [12][15][17]-[19][21]-[23] reported concerning levels of multidrug resistance (MDR). The overall prevalence of multidrug resistance was 53.5% (95% CI: 50.8% - 56.1%). The coagulase-negative staphylococci (CoNS) and Klebsiellapneumoniae displayed the highest MDR rates, reaching 63.5% (95% CI: 58.7% - 68.1%) and 76% (95% CI: 56.6% - 88.5%), respectively. Escherichiacoli (60.6%) and Staphylococcusaureus (54.7%) also showed significant resistance (Table 4).

Table 4. Distribution of multidrug resistance profiles among bacterial isolates from ocular infections in 8 African studies.

*CoNS: Coagulase-negative staphylococci; MDR: Multidrug-resistant (resistant to ≥3 antibiotic classes); R0 - R5: Number of antibiotic classes to which isolates are resistant.

Egger’s regression test revealed no significant funnel plot asymmetry (p > 0.05), and the symmetrical distribution of studies around the pooled estimate suggests no major publication bias. Trim-and-fill analysis confirmed the robustness of the findings, with an adjusted estimate nearly identical to the original. However, the limited number of studies (n = 15) reduces the power of these tests, although the comprehensive search strategy minimizes the risk of bias (Figure 8).

Figure 8. Funnel plot of ocular infection prevalence for publication bias assessment.