Bacterial profile and antimicrobial resistance patterns of common bacteria among pregnant women with bacteriuria in Ethiopia: a systematic review and meta-analysis

Background Globally, antimicrobial resistance (AMR) restricted the armamentarium of the health care providers against infectious diseases, mainly due to the emergence of multidrug resistant. This review is aimed at providing contemporary bacterial profile and antimicrobial resistance pattern among pregnant women with significant bacteriuria. Methods Electronic biomedical databases and indexing services such as PubMed/MEDLINE, Web of Science, EMBASE and Google Scholar were searched. Original records of research articles, available online from 2008 to 2021, addressing the prevalence of significant bacteriuria and AMR pattern among pregnant women and written in English were identified and screened. The relevant data were extracted from included studies using a format prepared in Microsoft Excel and exported to STATA 14.0 software for the outcome measure analyses and subgrouping. Results The data of 5894 urine samples from 20 included studies conducted in 8 regions of the country were pooled. The overall pooled estimate of bacteriuria was 15% (95% CI 13–17%, I2 = 77.94%, p < 0.001) with substantial heterogeneity. The pooled estimate of Escherichia coli recovered from isolates of 896 urine samples was 41% (95% CI 38–45%) followed by coagulase-negative Staphylococci, 22% (95% CI 18–26%), Staphylococcus aureus, 15% (95% CI 12–18%), Staphylococcus saprophytic, 12% (95% CI 6–18%) Proteus mirabilis, 7% (95% CI 4–10%), Enterococcus species, 6% (0–12%), Pseudomonas aeruginosa, 4% (2–6%), Citrobacter species, 4% (95% CI 2–4%), Group B streptococcus, 3% (1–5%), and Enterobacter species, 2% (1–4%). Multidrug resistance proportions of E. coli, Klebsiella species, Staphylococcus aureus and Coagulase negative staphylococci, 83% (95% CI 76–91%), 78% (95% CI 66–90%), 89% (95% CI 83–96%), and 78% (95% CI 67–88%), respectively. Conclusion The result of current review revealed the occurrence of substantial bacteriuria among pregnant women in Ethiopia. Resistance among common bacteria (E. coli, Klebsiella species, Staphylococci species) causing UTIs in pregnant women is widespread to commonly used antibiotics. The high rate of drug resistance in turn warrants the need for regular epidemiological surveillance of antibiotic resistance and implementation of an efficient infection control and stewardship program. Supplementary Information The online version contains supplementary material available at 10.1007/s00404-021-06365-4.


Introduction
Urinary tract infection (UTI) is referred to the invasion of microorganism and their subsequent growth in any part of the urinary tract including the kidneys, ureters, bladder, and urethra [1]. It is one of the most common infectious disease resulting in over 150 million cases per year globally [1,2]. Although both males and females may become infected by UTI, females are more likely to develop UTI compared to males due to short urethra, proximity to anus which causes easy contamination of the urinary tract with fecal microbial, hormonal changes and pregnancy [3].
The most frequent bacterial infection during pregnancy is urinary tract infection, which is linked to preeclampsia, low birth weight, premature delivery, and intrauterine growth restriction [4]. UTIs are classified as either asymptomatic or symptomatic bacteriuria during pregnancy. The presence of a significant amount of bacteria (10 5 colony-forming unit (CFU)/ml of cultured urine) in the absence of signs and symptoms of UTI is known as asymptomatic bacteriuria (ABU) [5,6]. If left untreated, asymptomatic bacteriuria occurs 2-15% of pregnant women and is a common risk factor for pyelonephritis [7,8]. Although, recent studies question the benefit of ABU treatment particularly, in low-risk pregnant women [7][8][9], timely initiation of antibiotic as soon as significant bacteriuria is diagnosed likely reduce the incidence of pyelonephritis, preterm delivery, and low birth weight [6]. Several factors are associated with an increased frequency of bacteriuria during pregnancy. These will include age, presence of genitourinary abnormalities (kidney, ureteral and bladder stones, tumors, urethral strictures, vesico-ureteric reflux), anemia, sexual activity, decreased immunity, lower socioeconomic classes, past history of UTI, multiparity and gestational age [10,11]. For example, the prevalence for bacteriuria increases with age from about 1% in females aged 5-14 years [12].
Escherichia coli is the most common pathogenic microorganisms associated with both asymptomatic and symptomatic bacteriuria and accounting up 60-80% of all UTIs in pregnancy [13,14]. Other bacteria identified in UTIs during pregnancy include Klebsiella pneumoniae, Staphylococcus aureus, Coagulase negative staphylococci, Enterobacter spp., Pseudomonas aeruginosa, Enterococcus spp., Proteus mirabilis, and others [14].
Antimicrobial resistance (AMR) continued to increase as a result of the rapid emergence of resistant bacteria worldwide [15]. This is demonstrated by recent world health organization (WHO) surveillance data from 22 high-and low-income countries which reported a high level of AMR to numerous bacterial infections and identified E. coli and K. pneumoniae as the most common resistant pathogen [16]. In addition, meta-analysis of 23 studies from different part of the world reported a high prevalence of extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae in pregnant women with significant bacteriuria [17]. The resistance of uropathogenic to thirdgeneration cephalosporin, amoxicillin, and other antibiotics in pregnant women especially, in Africa pose a big challenge to the management of UTI during pregnancy [17,18].
Routine antimicrobial susceptibility test is a major challenge in the majority of low-income countries [19]. As a result, treatment of infectious disease including UTI in pregnant women is initiated empirically and most pregnant women treated with an unnecessary antibiotic which further contributes to the emergence of resistant pathogens [20]. In Ethiopia, there are several studies conducted at different regions and hospitals regarding the bacterial profile and antimicrobial susceptibility patterns of UTI in pregnant women [21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40]. However, there is no pooled prevalence data at the national level which alert health care providers and clinicians on the current prevalence of AMR in UTI during pregnancy. Therefore, this systematic review and metaanalysis is aimed at providing contemporary bacterial profile and antimicrobial resistance pattern among pregnant women with significant bacteriuria.

Reporting
This systematic review and meta-analysis is based on a recommended methodology by the Joanna Briggs Institute's approach [41] and followed the Preferred Reporting Items for Systematic review and Meta-Analysis for Protocols (PRISMA-P) 2015 guidelines [42].

Search strategy and data extraction
The present review exhaust all available full-text articles in English language (without date restrictions) indexed in the following major scientific databases: MEDLINE (via PubMed), Web of Science (via Saudi Digital Library), Excerpta Medica dataBASE (EMBASE) and Google scholar. To identify potential search terms and retrieve the best set of results possible, a pre-determined MED-LINE search strategy was constructed using the PICO (patients, interventions, comparators, outcomes) equivalents: patients, exposures of interests, comparison, and outcomes. Accordingly the PECO framework used in our search was described as follows: population (P): pregnant women with bacteriuria in Ethiopia; exposure (E): significant bacteriuria; comparison (C): pregnant women without significant bacteriuria in Ethiopia; outcome (O): bacteria profile, antibiotic resistance profile.
A manual search for additional relevant studies using references from retrieved articles and related systematic reviews was also performed to identify original articles we might have missed. In addition, to find unpublished papers relevant to this systematic review and metaanalysis, some research centers and digital library of higher institutions found in Ethiopia was also explored. The search strings or terms were stemmed from the following keywords: pregnant women, pregnant mother, pregnancy, bacteriuria, urinary tract infection, UTI, significant bacteriuria, uropathogens, antibiotic resistance, antimicrobial resistance, multidrug resistance and Ethiopia. In the advanced searching databases, the searching strategy was built based on the above-mentioned terms using the "Medical Subject Headings (MeSH)" and "All fields" by linking "AND" and "OR" Boolean operator terms as appropriate. The pre-determined MEDLINE search strategy was adopted in searching relevant articles in all other databases. Modifications to indexing terms (e.g., Medical Subject Headings, field tags) for other databases were applied as necessary.

Data extraction and quality assessment
Data extraction was implemented using a standard and extraction format adapted from the Joanna Briggs Institute (JBI) data extraction format [41]. Screening by title, abstract, and full text and data extraction were done independently by two authors (LC and TM) from the published studies and recorded by Microsoft excel spreadsheet. Consensus was reached by discussion whenever there was disagreement. References and data for each study were carefully cross-checked to ensure that no overlapping data were present and to maintain the integrity of the meta-analysis. Endnote citation manager software version X9 for Windows was utilized to collect and organize search outcomes and for removal of duplicate articles. The retrieved articles were screened according to predefined inclusion and exclusion criteria. We extracted information on name of the first author and year of publication, study period, study design, study region, total sample size, number of isolates, criteria for diagnosis of significant bacteriuria, number and percentage of gram positive and gram negative bacteria, number of asymptomatic and symptomatic bacteriuria and the prevalence and antimicrobial resistance patterns of common bacteria identified. The JBI quality appraisal criteria established for cross-sectional were used to measure the quality of the studies included in the meta-analysis. The JBI's appraisal criteria consist of nine questions. The first three questions measure about appropriateness of the sampling frame to address the target population, whether participants sampled in appropriate way and sample size is adequate enough. The next three questions measure whether the study subjects and setting of the study described in detail, the data analysis conducted with sufficient coverage of the identified samples and validated for identification method of the condition. The last three questions measure the reliability of measuring the condition in a standard way, the appropriateness of statistical analysis and the adequacy of response rate.

Inclusion and exclusion criteria
We included published and unpublished (Addis Ababa University) studies reported bacterial profile and their antimicrobial-resistance in pregnant women. We also included studies published in English, human studies and conducted only in Ethiopia. We excluded studies in other population other than in pregnancy, qualitative studies, review paper, commentaries, case series, case reports, conference proceeding and abstracts.

Data processing and analysis
Data were extracted in Microsoft Excel format, followed by analysis using STATA Version 14.0 statistical software. Random effect model was applied to estimate the pooled estimate and antimicrobial resistance pattern of the isolates. We conducted meta-regression to understand the source of heterogeneity and pooled the estimate using "metaprop" command. Potential source of heterogeneity was investigated by subgroup and meta-regression analysis. The existence of heterogeneity among studies were examined by I 2 heterogeneity test, in which 0-40%, 50-60%, 50-90% and 75-100% represented low, moderate, substantial and considerable heterogeneity, respectively. I 2 heterogeneity test of ≥ 50% and a p value of < 0.05 was assured the presence of heterogeneity. Thus, the DerSimonian-Laired random effects model was employed [43]. Subgroup analysis was done based on the patients symptoms status (asymptomatic vs. symptomatic), bacterial isolates (Gram positive vs. Gram negative), geographical region, study period (before 2015 vs. 2015 or after) to minimize the random variations between the point estimates of the primary study. Visual inspection of the funnel plots and Egger's regression test was used for evaluating the possibility of publication bias. Due to the presence of publication bias the result was corrected by Duval and Tweedie's trim and fill analysis. Forest plot format was used to present the pooled estimate with 95% CI. Two-sided p values < 0.05 were accepted as statistical significant.

Outcome measurement
This review has three major outcomes. The first outcome was to determine the pooled estimates of significant bacteriuria among pregnant women in Ethiopia. The second outcome was to assess the rate of antimicrobial resistance of common bacteria to common antibiotics used in the country for the treatment of urinary tract infection and finally pattern of multidrug resistance of common uropathogens among pregnant women. The pooled estimates of bacterial isolates and their antimicrobial resistance isolates were calculated using the following formula: a pooled estimate of each isolate was the number of isolate divided by the number of all the detected bacteria isolates. The pooled estimates of resistance to each tested antibiotic was calculated by dividing the number of resistance isolates of each species to total number of all the detected isolates of the species. Multidrug resistance was defined as resistance to at least to two antibiotics to the isolated Gram-positive bacteria for the various antibiotics.

Pooled estimate of significant bacteriuria
A total of 20 studies reported that significant bacteriuria was detected in 896 urine samples out of a total 5894 samples taken from pregnant women during their antenatal care visit to the health care system. In random-effect model, the overall prevalence of significant bacteriuria was 15% (95% CI 13-17%) with substantial heterogeneity (I 2 = 77.94%, p < 0.001) across the studies (Fig. 2).

Meta-regression
Just to identify the sources of heterogeneity, meta-regression was conducted using geographic region, year of publication, and sample size as a covariate. The analysis showed that there is no effect of geographic region (supplementary 1), year of publication (supplementary 2) and sample size (supplementary 3) on heterogeneity between studies ( Table 2).

Publication bias
The asymmetry of funnel plots visual inspection showed that there are some evidence of publication biases in the pooled estimates, which were statistically confirmed by Egger's regression test (β = 8.30, 95% CI 4.30-12.30, p < 0.001, Supplementary 11) and after adjusting for publication bias by trim and fill analysis the funnel plot appeared symmetrical (Supplementary 12).

Prevalence
In this meta-analysis of 20 studies, 15% of the pregnant women attending the antenatal care follow at different regions of Ethiopia had significant bacteriuria. Our finding showed a slightly higher number of bacteria identified compared to a review study from other lowincome countries, where 13.5% of UTI-causing bacterial uropathogens were isolated from the urine sample [14].  In contrast to our meta-analysis, this review included a larger sample size (24,248 vs. 5536 urine specimens), however, the prevalence estimate did not contain the rigorous statistical analysis including the sub-group analysis. However, from a multicentre study of Tanzania, a slightly higher prevalence of significant bacteriuria among pregnant women was reported (17.7%); compared to our finding [19]. In different studies, Greve et al. [53], Lee et al. [54] and Prifti et al. [55], lower findings were reported, where the urine culture positivity (bacteriuria) among 8807, 4034, and 2149 pregnant women were 5.6%, 8.9%, and 13.31%, respectively. Compared with the rest of the world especially the developed countries, the prevalence of bacteriuria among pregnant women in Ethiopia were high. These could be associated with the different modifiable and preventable prenatal factors. These will include anemia, decreased immunity, sexual activity, early age marriage, lower socioeconomic classes, multiparity and gestational age [56][57][58][59][60]. Untreated asymptomatic bacteriuria in pregnancy leads, in as much as 40%, to the development of acute pyelonephritis with all the subsequent negative effects not only for the woman herself but particularly for the foetus [52]. This meta-analytic result indicated 12% of the pregnant women with significant bacteriuria were reported to be asymptomatic. A similar meta-analytic result from Iran indicated 13% of pregnant women had asymptomatic bacteriuria [61]. However, our finding is higher compared to a study by Lee et al. (Bangladesh), where 4.5% of the pregnant women with UTI had asymptomatic bacteriuria [54] and lower compared to a study by Mwambete and Malaba [62] from Tanzania, Dar es Salaam where 23.3% of the pregnant women attending the antenatal care had an asymptomatic urinary tract infection.

Strengths and limitations
The main strength of this meta-analysis is that it was the first to be reported from Ethiopia. In addition, it pooled a relatively large number of articles, which could improve the power and the precision of the estimates of the effect sizes. While the study findings led to the conclusion of high antibiotic resistance in pregnant women with UTI, the original studies included in the analyses came from only six of the country's nine regions and two cities. As a result, the results' applicability to the rest of the country may be questioned. In this meta-analysis, there is also significant publication bias, which is graphically assessed using funnel plots and statistically checked for the presence of a small study effect using the Egger test, affecting the interpretation of the findings. However, the pooled prevalence was corrected by Duval and Tweedie's trim and fill analysis (the same as the original result, 15%). Furthermore, the reader should be aware that the protocol for this study was not made public, which could lead to bias.

Conclusion
The result of current review revealed the occurrence of substantial bacteriuria among pregnant women in the country. Resistance among common bacteria (E. coli, Klebsiella species, Staphylococcus species) causing UTIs in pregnant women is widespread to commonly used antibiotics. The high rate of bacteriuria among pregnant women in the country is an alert for clinicians to screen bacteriuria at least once using urine culture during antenatal care and treat if urine culture turn around positive. The high rate of drug resistance in turn warrant the need for regular epidemiological surveillance of antibiotic resistance and implementation of an efficient infection control and stewardship program.
Author contributions LCK and TM conducted the database search, screening, quality assessment, and data analysis. KF and TMU wrote and revised the manuscript for the intellectual content. All authors read and approved the final manuscript.
Funding The authors received no specific funding for this work.
Availability of data and materials All relevant data are within the manuscript and its Supporting Information files.

Conflict of interest
The authors declare that they have no competing interests.
Ethics approval and consent to participate Not applicable.

Consent for publication Not applicable.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.