Prevalence of vancomycin-resistant enterococcus in Africa in one health approach: a systematic review and meta-analysis

Vancomycin-resistant enterococci are a global challenge currently as reported by the World Health Organization. It is also important to recognize that combating antimicrobial resistance needs to recognize the interconnections between people, animals, plants and their shared environment in creating public health, the so-called One Health approach. Although the presence of VRE has been described in many regions of the world, there is a lack of comprehensive data indicating their prevalence of in Africa. Therefore, this study aimed to aggregate the result of studies describing VRE reported across multiple regions in Africa. A literature search was conducted on PubMed, Google scholar, and Hinari with the term “Vancomycin resistance enterococcus in Africa” on August 1–3, 2019. All available articles were downloaded to “Endnote version 7.1” then to Microsoft Word 2013. Articles determined to meet our criteria for the review was extracted to Microsoft Excel 2013. Those articles that reported the prevalence of vancomycin resistance Enterococcus obtained from all sample types and published from 2010 to 2019 in the English language were included for the review. A meta-analysis was conducted with OpenMetaAnalyst version R.3.1.0 software. The effect size was determined using a binary random effect model and statically significant considered when p < 0.05. Heterogeneity determined with the inconsistency index. A leave one out analysis used to perform the sensitivity analysis. There were 151 articles identified from the database searches; of this, 36 articles included after extensive review with two independent authors. Out of 4073 samples collected, 1488 isolates identified with an overall pooled prevalence of VRE 26.8% (95% CI; 10.7–43.0%) in Africa with a one-health perspective. The analysis showed that considerable heterogeneity among the studies (I2 = 99.97%; p < 0.001). Subgroup analysis in-country, African region, laboratory method, year of publication, and sample source showed that a high prevalence was identified from South Africa (74.8%), South African regions (74.8%), PCR (959.2%), 2010–2015 years (30.3%) and environmental (52.2%), respectively. This meta-analysis indicates that there was a high-pooled prevalence of vancomycin-resistant enterococci in African. A lot should be done to prevent and control the transmission of vancomycin resistance enterococci to a human being from the environment in the continent.

www.nature.com/scientificreports/ to reptiles, birds, and mammals. Whilst they are ubiquitous, they represent a minority population of the healthy human microbiome 6 . Presence in the environment is generally considered an indicator of human or animal faecal contamination of recreational or drinking water 7 . The rise of VR Enterococcus feacium in the European Union has to lead to the sanction of avoparcin, an antibiotic that chemically related to vancomycin 8 . The USA never approved avoparcin for clinical use. In the years post-ban, VRE surveillance data of EU hospitals showed no obvious reduction in VRE rates. Because of very limited alternatives to vancomycin, VRE infections remain a serious clinical treatment challenge throughout the world. Surveillance studies showed zero rates of VRE in US livestock. Whole-genome sequencing data suggest that VRE might have evolved from ampicillin-resistant E. feacium from dogs 9 .
Some members of the genus Enterococcus are well-documented pathogens associated with serious clinical manifestations in humans, including bacteremia, infective endocarditis, intra-abdominal and pelvic infections, urinary tract infections, and, in rare cases, central nervous system infections [10][11][12] . Infection with VREs is associated with an increased mortality rate, illustrated by a 2.5-fold increase in mortality for patients suffering from VRE bacteremia 13 .
The One Health Commission defines One Health as "a collaborative, multisectoral, and trans-disciplinary approach-working at local, regional, national, and global levels-to achieve optimal health and well-being outcomes recognizing the interconnections between people, animals, plants and their shared environment." All potentially constitute overlapping reservoirs of antimicrobial resistance 14 . Given the serious health threat, a common understanding of AMR, and of the need for a One Health approach to tackle it, are of fundamental importance 15,16 .
VRE is one of these multidrug resistances that need comprehensive data that indicates the pooled prevalence of VRE in Africa. Therefore, this study aimed to compile available data of VRE in Africa in a one-health perspective: a systematic review and meta-analysis.

Methods
Literature search strategy. A literature search conducted on PubMed, Google scholar, and Hinari with the term "Vancomycin resistance enterococcus in Africa" on August 1-3, 2019. Citations of all available articles were exported to "Endnote version 7.1" then to Microsoft Word 2013. All the articles that met our inclusion criteria were included for systematic review and meta-analysis. There were 151 articles obtained from the databases. Of these, 29 articles were excluded based on setting and duplications, 66 were excluded because either title or year of publication was unacceptable. A total of 56 articles underwent full-text assessment. An additional 20 articles were excluded because they failed to report the prevalence of VRE, their year of publication was before 2010, or they lacked clearly described laboratory methods or unknown sample source. Finally, 36 articles were subjected to an extensive review by two independent authors. The article selection process was conducted according to the PRISMA protocol of 2015 17 (Fig. 1).
Eligibility criteria. The inclusion criteria for this review were articles published in the English language that reported the prevalence of VRE and were published from 2010 through 2019. All sample sources were considered. Specifically, publications excluded from this review were any of the following: published before 2010 or after 2019, published other than the English language, has no clear laboratory methods, had unknown sample types, or failed to include the prevalence of VRE. Data analysis. Microsoft Excel 2013 was used for data extraction and results were then exported to Microsoft Word plus 2013. Data was entered to OpenMetaAnalyst version R.3.1.0 18 software for each study after copying each column from Excel to the software and a meta-analysis, subgroup meta-analysis and sensitivity analysis were conducted. The result was presented as a forest plot in the figure. The pooled prevalence of VRE at 95% CI was determined with a binary random-effect model by the DerSimonian-Laird method. The statistical significance was considered when p value < 0.05. Data quality. The quality of the study included in the review and meta-analysis evaluated with a 14 point scoring tool, an NIH quality assessment tool for observational and cross-sectional studies in which studies categorized as a good, fair, and poor quality based on the internal validity of each article 19 . Accordingly, nine (25%) articles were categorized as fair, eleven (30.6%) as poor, and sixteen (44.4%) articles as good quality (supplementary file).
Heterogeneity and publication bias. The heterogeneity of the publication was determined with the measure of the inconsistency index (I 2 ) and p value. The total variations in the articles were due to heterogeneity rather than by chance with a value of < 30%, 30-60%, 61-75%, and > 75% suggestive of low, moderate, substantial, and considerable heterogeneity, respectively 20 . Publication bias was not checked as the study is considerably heterogeneous as recommended by Hak et al. 21 , if the data is heterogeneous no need of conducting publication bias.

Study features.
Studies conducted in African countries that reported the prevalence of VRE and were published between 2010 and 2019 were considered. All sample types and laboratory methods employed were included for the review and meta-analysis. The following data types were extracted from each article and presented in Table 1 Country of origin for the articles. The country in which the articles originated is indicated as follows, eight articles from each country, Ethiopia [22][23][24][25][26][27][28][29] , and South Africa 29-37 , four articles in each country Egypt [38][39][40][41] , and Tunisia [42][43][44][45] . Another three articles from each of these countries Morocco [46][47][48] and Uganda 49-51 and two articles from each of these countries Nigeria 52,53 , Tanzania 54,55 and Algeria 56,57 were included for the study (Table 1).

Discussions
Vancomycin is one of a limited number of antibiotics that can be used to treat infections in humans resulting from Gram-positive multidrug-resistant organisms (MDRO) including Enterococci 58 . In the late 1980s, the emergence of VRE in European hospitals followed by isolation from Danish raw minced pork and frozen poultry generated global concern 59 . One Health is the concept that the optimum health for people, animals, and the environment should all be considered through the ongoing cooperative efforts of scientists and practitioners in a variety of disciplines 60 .
Our study based on the data available from studies in Africa on VRE in which animal, human, and environmental sources of samples had been specified were analyzed to determine the pooled prevalence of VRE. The overall prevalence of VRE was (26.8%) in Africa from different sample sources. This prevalence is higher than reported in the studies conducted in Iraq (14%) 61 , Europe (2.7%) 1 , (13%) 62 , Thailand (10.3%) 63 , South America (6%) 64 . However, it is comparable to a study from Latin America (30%) 65 . These differences may be due to the source of the sample we used for the analysis is from different sources and may be due to the enterococcus population structure selected overtime which is highly resistant for environmental conditions and different antibiotics 2,66 .
The subgroup analysis at the country level showed that there is a pronounced difference of VRE in different countries, which ranged from (74.8%) in South Africa to (2.8%) in Algeria and Nigeria which is statically significant with p < 0.000. This variation might be due to sampling source difference, sample size, laboratory method used, year of publication, and the number of studies included for the meta-analysis.
Our study also performed a subgroup meta-analysis based on the laboratory method used for isolation and identification of VRE. It showed that the higher the technique engaged by the studies for identification of VRE, the more sensitive and specific for detection of VRE in which studies conducted with PCR primers for isolation of higher prevalence of VRE, whereas those conducted with conventional culture were less likely to detect VRE. Some studies reported in a comparison of PCR and culture supports that the former technique is more sensitive and specific than later one for the identification of VRE [67][68][69] .   70 . The disagreement might be due to study period variance, the area covered, sample type used, the ability of detection of laboratories dissimilarity.
Analysis of VRE in African regions showed that there was a high prevalence in the South African region (74.8%) almost twenty-six times of West Africa and four times than of North and East African regions. The difference can be explained it might be due to the laboratory method used for detection and identification of VRE 67,69 , the sample difference 71 and overall antibiotics usage in human 72,73 and animal [74][75][76] .
The sample source in which we categorized in human, animal and environmental sources for the sake of subgroup meta-analysis showed that a higher prevalence of VRE was isolated from environmental, followed by the animal source as compared to a human source. This may be due to most of the articles included based on our inclusion criteria is from non-human sources as different wild and domestic animal wastes or by-products, poultry, birds, and the environmental sample was compiled for analysis. The other reason is probably due to the intensive conditions in which these animals maintained for different antibiotics as a kind of growth promoter 77,78 . This part of the study strained the one health approach, which is an important way of combating antibiotics resistance that worsens the world; now a day's 77,79 .

Conclusion
Our meta-analysis finding demonstrated that there is a high prevalence of VRE circulating in Africa. The subgroup analysis indicates that a high prevalence of VRE isolated from South African region. Similarly, studies conducted with PCR laboratory method isolated the highest VRE. Additionally, our study showed that the prevalence decreasing over time. Environmental sample source is with a higher VRE as compared to the human and animal sample source. Thus, a means of prevention and control targeting humans, animals, and environments based on regional, and country perspectives should be practised in the continent to alleviate the infection with VRE.

Data availability
All the data supporting the findings can be obtained from the corresponding author.  www.nature.com/scientificreports/