Prevalence and pathogen profiles of bacteremia in neonates hospitalized for clinical Sepsis in Ethiopia: a systematic review and meta-analysis Authors: Daniel Geleta, Gemeda Abebe, Tsion Tilahun, Hunde Ahmed, Netsanet Workneh, Getenet Beyene Published: 2024 ABSTRACT BackgroundBacteremia is prevalent in neonates, largely attributed to factors inherent in the neonatal period. However, the prevalence of proven bacteremia in Ethiopian neonates has not been previously synthesized. Accordingly, this systematic review and meta-analysis aimed to analyze the prevalence of bacteremia and pathogen profiles in neonates hospitalized for clinical sepsis in Ethiopia.MethodsThis systematic review and meta-analysis followed the preferred reporting items for systematic review and meta-analysis (PRISMA) 2020. The literature search was conducted across five databases including PubMed, Google Scholar, Web of Science, Science Direct, and Research for Life spanning from January 2015 to July 2023. The search strategy used MeSH terms and involved screening titles and abstracts, reviewing full-text articles, and including only observational studies published in English within the specified timeframe. Data extraction and quality assessment were performed by three experienced reviewers using a validated data collection tool and the Joanna Briggs Institute quality assessment tool, respectively. The prevalence of neonatal bacteremia was determined through a random effects model, with heterogeneity among studies assessed using the Q statistic and the I2 statistic. Publication bias was evaluated using a funnel plot and Egger’s regression test, and STATA version 16.0 was used for all analysis at 95% confidence level.ResultsA meta-analysis of nine studies revealed a bacteremia prevalence of 40.0% (95% CI: 34.0–46.0%). Subgroup analysis indicated variations in prevalence based on regions and study designs, with Oromia at 44.0% (95% CI: 28.0, 61.0%) and Amhara at 39.0% (95% CI: 27.0, 51.0%). Longitudinal studies exhibited a higher prevalence (47.0%, 95% CI: 27.0–68.0%) compared to cross-sectional designs (38.0%, 95% CI: 32.0–44.0%). Gram-negative bacteria were identified as the predominant etiological agents, representing 59.5% (95% CI: 56.8–62.3%) of cases. Among the bacterial species, Staphylococcus aureus emerged as the most prevalent (20.0%, 95% CI: 18.0%, 22.0%), followed by coagulase-negative staphylococci and Klebsiella pneumoniae, each contributing to 17.0% (95% CI: 15.0%, 20.0%) of bacteremia cases.ConclusionThe study revealed a significant high prevalence of bacteremia, with differences noted across regions and study designs. Key pathogens identified were Staphylococcus aureus, Klebsiella pneumoniae and coagulase-negative staphylococci. It is advisable to implement surveillance systems, targeted prevention strategies, diagnostic stewardship, and further research on regional variations and bacterial profiles to effectively enhance the ominous future.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12879-024-10312-4. Introduction Neonatal bacteremia, a serious condition characterized by the presence of bacteria in the bloodstream of neonates, poses significant challenges in accurately determining its prevalence [1]. The complexity commonly stems from incomplete birth and death registration, hindering comprehensive data collection and reporting [2]. Even in cases with complete registration, accurately identifying the specific bacteria causing the infection necessitates time-consuming and labor-intensive blood culture, which is considered the most reliable diagnostic method [3]. Indeed, globally, and especially in developing regions, neonatal bacteremia is a significant public health issue that often receives inadequate attention [4]. In these settings, limited availability of laboratory diagnoses and a lack of strong diagnostic stewardship contribute to the challenges in accurately identifying and managing neonatal bacteremia [2, 5]. The scarcity of laboratory resources hampers the ability to conduct timely and reliable diagnostic tests [6]. As a result, healthcare providers heavily rely on clinical algorithms that prioritize sensitivity over specificity [3, 7]. Consequently, the available epidemiological data on neonatal bacteremia are fragile and show significant variation across different locations, healthcare settings, and populations studied [4, 8]. The occurrence of bacteremia is higher in neonates compared to other age groups due to unique characteristics present during the neonatal period [4]. These characteristics include exposure to multiple iatrogenic risks, an immature immune system [3], the inextricable link with maternal perinatal care and well-being [9], the lack of point-of-care diagnostic tests [10], disparities in health care quality [11, 12], institutional culture and climate, and heterogeneity of the neonatal population, all of which play significant clinical importance [13, 14]. Numerous studies have investigated the prevalence of neonatal bacteremia, revealing rates that range from less than 1% to over 10% [15]. Particularly, a systematic review and meta-analysis found that the global pooled prevalence of neonatal bacteremia was 2.01% [4]. Specifically, the prevalence was 2.9% in the United States [15], 8.2% in Pakistan [16], 20.5% in Nepal [17], 37.6% in India [18], 47% in Nepal [19] and 64/8% in Iraq [20], with no pooled bacteremia prevalence data in Ethiopia to the extent of this review. Further, these infections are commonly attributed to both gram-positive and gram-negative bacteria (53%) [4]. In the United States, gram-positive pathogens, such as Staphylococcus aureus and coagulase-negative staphylococci, are frequently identified in neonatal bloodstream infections, especially in neonatal intensive care units (NICUs) [21]. Coagulase-negative staphylococci, like Staphylococcus epidermidis, have a tendency to colonize medical devices, thereby increasing the risk of infection. Similarly, gram-negative pathogens like Escherichia coli and Klebsiella spp. are also associated with neonatal bacteremia. Escherichia coli commonly causes early-onset sepsis, often linked to maternal colonization and ascending infections, while Klebsiella pneumoniae can cause both early and late-onset sepsis [4, 21]. It is important to acknowledge that the epidemiological characteristics of gram-positive and gram-negative pathogens in Ethiopia may differ from those observed in other developing and developed countries [22]. Factors such as variations in prevailing bacterial strains, antimicrobial resistance patterns, healthcare practices, and host factors can contribute to these dissimilarities [21]. Given the different epidemiologic characteristics of gram-positive and gram-negative pathogens in Ethiopia compared to other countries, previous studies have conducted systematic reviews and meta-analyses specifically focusing on the clinical prevalence and predominant pathogen of neonatal bacteremia. Consequently, the objective of this review was to synthesize the current evidence, pinpoint gaps in research, evaluate the quality of available evidence, produce summary estimates, and provide guidance for both clinical practice and policy-making within the Ethiopian context regarding blood culture-proven bacteremia. Methods Study design A systematic review and meta-analysis, following the preferred reporting items for systematic review and meta-analysis (PRISMA) 2020 [23] guidelines, were conducted, with registration in PROSPERO under the number CRD42023447381 for international prospective review tracking. Search strategy and selection criteria PubMed, Google Scholar, Web of Science, Science Direct, and Research for Life were searched for potential articles reporting the prevalence of blood culture proven neonatal sepsis in Ethiopia. The search was limited to articles published in English between January 2015 and July 2023 because most health research studies conducted in Ethiopia published using the English language. In addition to the electronic databases, a hand search was conducted for grey literature. The search terms included “prevalence”, “early-onset sepsis”, “late-onset sepsis”, “bacterial isolates”, “bacterial pathogen”, “bacteria”, “bloodstream infection” “bacteremia,” “sepsis” “septicemia” “antibiotic resistance pattern”, “antimicrobial resistance”, “antimicrobial susceptibility” “antimicrobial susceptibility test” “neonates”, “newborn”, “Neonatal”, “young infants”, “ neonatal intensive care unit”, “NICU” and “Ethiopia”. The search terms / Medical Subject Headings (MeSH) were further combined with Boolean operators like “AND” and “OR” to narrow or broaden the search strategy, respectively. Only observational studies of original nature, reporting information on the prevalence of blood culture-proven bacteremia according to the Clinical and Laboratory Standards Institute 2023 guidelines of blood culture [24], and including the number of neonates with sepsis were included. Studies with single bacterial prevalence detection, study settings outside of NICU, and studies involving adult population subjects were excluded (Supplementary Table 1). Outcome variables, data selection and extraction The proportion of neonate proven bacterial sepsis to the total number of collected blood samples, sex, and type of sepsis, and causative etiologies were considered the outcomes of the study. After duplicate removal of articles using Mendeley reference manager and screening of title and abstract, articles were rigorously reviewed, and data collection was conducted independently by DG, HA and TT. Data were extracted from the selected articles and compiled using Microsoft Excel spreadsheets that prepared by the review team during protocol development. The spreadsheet contained name of primary author, publication year, study pried, sample size, sex of neonates, number of blood culture positive samples, number of diagnosed neonates for sepsis and place of study. The consistency of the extracted data was verified and approved by the other authors (GA, NW, and GB) to ensure accuracy and reliability. Study quality appraisal and risk of bias assessment The authors employed the Newcastle-Ottawa Quality Assessment Tool scale [25] in conjunction with Joanna Briggs Institute (JBI) critical appraisal checklist for prevalence studies to rigorously assess the risk of bias in the included prevalence studies and ensure interrater agreement between reviewers [26]. This assessment tool comprises three components. The first component, rated on a five-star scale, evaluates the methodological quality of each study. The second component, rated on a three-star scale, assesses the comparability of the studies. The final component, rated on a two-star scale, measures the quality of the original articles in terms of their statistical analyses [27, 28]. The review included cross-sectional studies that met at least acceptable standards, fulfilling 50% of the quality assessment criteria, according to a comprehensive evaluation method. Inconsistencies were resolved through discussions and consensus among the authors, with the involvement of a third reviewer. The first author conducted the initial quality appraisal of the included studies, and then obtained validation from the second and third authors. Through this collaborative process, any disagreements or discrepancies were addressed and resolved to ensure the accuracy and reliability of the quality assessment. Likewise, for observational studies, an existing JBI quality assessment tool was utilized to evaluate the quality of the study. Data synthesis and meta-analyses The prevalence of proven bacteremia among neonates in Ethiopia was estimated using a random effects model, with a 95% confidence interval (CI). To visually represent the individual and pooled prevalence of the studies, as well as their corresponding 95% CIs, a forest plot was employed. The assessment of statistical heterogeneity in our research was conducted using the Q-statistic test and I2 statistic, with consideration of their corresponding 95% confidence intervals. The results were classified into different levels of heterogeneity, including low (≤ 25%), moderate (25 < I2 ≤ 50%), substantial (50 < I2 ≤ 75%), and considerable (75 < I2 ≤ 100%) [29]. Subgroup analyses were conducted based on region (administrative boundaries of the country), study design, type of sepsis and sex of participants. To assess reporting and publication bias, a funnel plot and Egger’s tests were used. Results In the initial web-based search, a total of 877 records were retrieved. However, after removing 498 duplicate records, there were 379 unique records remaining. Through a screening process that involved reviewing the titles and abstracts of 365 records, the selection was further narrowed down to 14 records. After conducting a full-text assessment of these 14 records, five studies were excluded for various reasons. These reasons included unclear study subject segregation [30–32], insufficient study criteria [33] and being a sister publication [34]. As a result, there were nine studies that met the eligibility criteria for inclusion in the systematic review. These nine eligible studies underwent a quality assessment, and all of them met the required standards [35–43]. Data extraction was then performed, and the findings from these studies were utilized in a meta-analysis and quantitative report (Fig. 1). Fig. 1PRISMA flowchart illustrating the search and study selection strategy PRISMA flowchart illustrating the search and study selection strategy Characteristics of the studies All the included studies were conducted in neonatal intensive care units (NICUs) of teaching hospitals between years 2017 and 2023 [35–43]. Except two [38, 43], all studies employed cross-sectional designs [35–37, 39–42], with three in Oromia [41–43], three in Amhara [36, 37, 39], and the rest in SNNP [35], Tigray [40], and multicenter [38]. In the total of, 4561 clinically sampled sepsis the largest sample size included 690 [38] neonates diagnosed with clinical sepsis, while the smallest included 246 [37] neonates. Of these clinically diagnoses sepsis cases, seven studies conducted blood culture tests for all suspected cases whereas two of the studies tested blood culture for 64.2% [43], and 34.5% [38] of cases. Only 33.3% of studies were provided bacteremia prevalence based on sex and age (type of sepsis) [35, 36, 39] (Table 1). Table 1Characteristics of studies included in the meta-analysis on blood culture-proven neonatal bacteremia prevalence in EthiopiaName of first author (Published year)Study DesignStudy periodRegion/SettingClinical sepsisTested blood samplesProven bacteremiaEOS bacteremiaLOS bacteremiaMale BacteremiaFemale BacteremiaTsehaynesh G et al. (2017) (36)Cross-sectionalSept 2015 to May 2016Amhara, University of Gondar Hospita (NICU)25125112087336849Abebe S et al. (2019) (41)Cross-sectionalApr 2016 to May 2017Oromia, Asella teaching and referral hospital (NICU)303303883652NANAWoldu Y et al. (2020) (40)Cross-sectionalMar 2017– Sept 2018Tigray, Ayder Comprehensive Specialized Hospital (NICU)3173171163185NANABeza E et al. (2020) (38)LongitudinalJul 2016 to May 2018Multicenter Tertiary Hospitals(NICU), AA, Amhara, Oromia2,003690255NANANANAMelkamu B et al. (2021) (43)Longitudinal, CohortApr 1 to Oct 31, 2018Oromia, Jimma Tertiary hospital (NICU)304188109NANANANATazeb M et al. (2021) (39)Cross-sectionalApr 2018 to Jul 2018Amhara, Felege-Hiwot Referral Hospital41241217090809080Endale W et al. (2022) (35)Cross-sectionalMar 2021 to Nov 2021Hawassa University Comprehensive Specialized Hospital(NICU), SNNRP39239214376678855Genet M et al. (2022) (37)Cross-sectionalAug 2017 to Mar 2018Amhara, Comprehensive Specialized Hospita(NICU)24624667NANANANATakele O et al. (2023) (42)Cross-sectionalOct 1, 2020 to Dec 31, 2020Oromia, Adama Hospital Medical College(NICU)3333171478562NANANote: NA represented note available Table 1Characteristics of studies included in the meta-analysis on blood culture-proven neonatal bacteremia prevalence in EthiopiaName of first author (Published year)Study DesignStudy periodRegion/SettingClinical sepsisTested blood samplesProven bacteremiaEOS bacteremiaLOS bacteremiaMale BacteremiaFemale BacteremiaTsehaynesh G et al. (2017) (36)Cross-sectionalSept 2015 to May 2016Amhara, University of Gondar Hospita (NICU)25125112087336849Abebe S et al. (2019) (41)Cross-sectionalApr 2016 to May 2017Oromia, Asella teaching and referral hospital (NICU)303303883652NANAWoldu Y et al. (2020) (40)Cross-sectionalMar 2017– Sept 2018Tigray, Ayder Comprehensive Specialized Hospital (NICU)3173171163185NANABeza E et al. (2020) (38)LongitudinalJul 2016 to May 2018Multicenter Tertiary Hospitals(NICU), AA, Amhara, Oromia2,003690255NANANANAMelkamu B et al. (2021) (43)Longitudinal, CohortApr 1 to Oct 31, 2018Oromia, Jimma Tertiary hospital (NICU)304188109NANANANATazeb M et al. (2021) (39)Cross-sectionalApr 2018 to Jul 2018Amhara, Felege-Hiwot Referral Hospital41241217090809080Endale W et al. (2022) (35)Cross-sectionalMar 2021 to Nov 2021Hawassa University Comprehensive Specialized Hospital(NICU), SNNRP39239214376678855Genet M et al. (2022) (37)Cross-sectionalAug 2017 to Mar 2018Amhara, Comprehensive Specialized Hospita(NICU)24624667NANANANATakele O et al. (2023) (42)Cross-sectionalOct 1, 2020 to Dec 31, 2020Oromia, Adama Hospital Medical College(NICU)3333171478562NANANote: NA represented note available Characteristics of studies included in the meta-analysis on blood culture-proven neonatal bacteremia prevalence in Ethiopia Note: NA represented note available Prevalence of bacteremia and heterogeneity The prevalence of bacteremia demonstrated substantial variation across the included studies, with rates ranging from 27.0% (95% CI: 22.0–33.0%) to 58.0% (95% CI: 51.0–65.0%). A meta-analysis incorporating all nine studies reported an overall pooled prevalence of 40.0% (95% CI: 34.0–46.0%). Notably, significant heterogeneity was observed among the studies, as indicated by an I² value of 92.19% (p < 0.01) in Fig. 2, suggesting that this variability is unlikely to have occurred by chance alone. A leave-one-out analysis was used to evaluate the influence of individual studies on the overall results, but no significant influence was found. Fig. 2Individual study effect size and pooled prevalence of neonatal bacteremia in Ethiopia Individual study effect size and pooled prevalence of neonatal bacteremia in Ethiopia The subgroup analysis found different rates of bacteremia in neonates based on several factors, including the geographical region and study design. According to Fig. 3, the data indicates that the Oromia region had the highest prevalence rate of bacteremia in neonates, with a rate of 44.0% (95% CI: 28.0, 61.0, I2 = 95.83%, P < 0.01). The second-highest prevalence rate was observed in the Amhara region at 39.0% (95% CI: 27.0, 51.0, I2 = 92.85%, P < 0.01), followed by the Tigray region with a rate of 37.0% (95% CI: 31.0, 42.0), and the Southern Nations, Nationalities, and Peoples’ Region (SNNRP) region with a rate of 36.0% (95% CI: 32.0, 41.0). The evidence clearly showed that there were meaningful differences in the prevalence of the bacteremia across Oromia and Amhara regions. The p-values derived from this testing showed the differences were statistically significant. Fig. 3Forest plot showing prevalence of neonatal bacteremia by region in Ethiopia Forest plot showing prevalence of neonatal bacteremia by region in Ethiopia The data presented in Fig. 4 indicates that the prevalence of bacteremia is influenced by the type of study design employed. Specifically, cross-sectional studies showed a relatively lower prevalence of bacteremia at 38.0% (95% CI: 32.0–44.0%), while longitudinal studies exhibited a higher prevalence at 47.0% (95% CI: 27.0–68.0%). It is important to note that the results of cross-sectional studies demonstrated significant heterogeneity (I2 = 88.2%, p < 0.01), indicating substantial variation in prevalence estimates ranging from 27.0% (95% CI: 22.0–33.0%) to 48.0% (95% CI: 42.0–54.0%). Conversely, longitudinal studies displayed a higher degree of heterogeneity (I2 = 96.3%, p < 0.01) with prevalence estimates ranging from 37% (95% CI: 33.0–41.0%) to 58% (95% CI: 51.0–65.0%). The evidence also supports a meaningful difference between cross-sectional and cohort study designs that occurred not due to chance alone. The corresponding statistical tests that were run on the data showed significant p-values, indicating it is highly unlikely that the differences seen between the study designs were simply due to random variation. Fig. 4Forest plot illustrating the prevalence of neonatal bacteremia by study design in Ethiopia Forest plot illustrating the prevalence of neonatal bacteremia by study design in Ethiopia Bacterial pathogens of neonatal bacteremia The analysis of the studies revealed that gram-negative bacteria were responsible for 59.5%( 95% CI: 56.8–62.3%) of cases, while gram-positive bacteria accounted for 40.5%( 95% CI: 37.8–43.1%). Among the identified pathogens, Staphylococcus aureus was the most commonly found, representing 20.0% (95% CI: 18.0, 22.0%) of cases. Coagulase-negative staphylococci and Klebsiella pneumoniae were equally prevalent, each accounting for 18.0% (95% CI: 15, 20.0.0%). On the other hand, Enterobacter cloacae was the least frequently identified pathogen, with a contribution of only 2.0% (95% CI: 1.0, 2.5%). Most of the identified pathogens were specified at the species level, with gram-negative bacterial species being the predominant pathogens (Fig. 5). Fig. 5Leading pathogens of proven neonatal bacteremia in Ethiopia Leading pathogens of proven neonatal bacteremia in Ethiopia Risk of bias within studies Both Egger’s test and the funnel plot analysis found no significant evidence of publication bias in the analysis of bacteremia prevalence. Egger’s test provided a statistically non-significant p-value (0.06). Additionally, the symmetric funnel plot depicted in Fig. 6 visually confirms the lack of publication bias through the even distribution of studies. Fig. 6Funnel plot analysis for bacteremia prevalence (symmetric) Funnel plot analysis for bacteremia prevalence (symmetric) Discussion A recent systematic review and meta-analysis of nine studies conducted in Ethiopia has revealed a substantial burden of bacterial sepsis among neonates. The analysis demonstrated that 40% of neonates presenting with clinical sepsis were confirmed to have bacterial pathogens. Notably, the prevalence of this condition exhibited statistically significant regional variations. The Oromia region reported the highest prevalence at 44.0%, followed by the Amhara region. Furthermore, the study design was found to significantly influence the observed prevalence, with longitudinal studies reporting a higher prevalence compared to cross-sectional studies. Gram-negative bacteria were the most commonly identified causative agents of bacteremia, although Staphylococcus aureus emerged as the primary pathogen, constituting 20.0% of the overall occurrence rate. The result of this meta-analysis conducted in Ethiopia revealed a higher pooled prevalence of proven neonatal bacteremia compared to rates reported in studies from Pakistan (8.2%) [16], Nepal (20.5%) [17] and United State (2.9%) [15]. However, the prevalence in Ethiopia showed similarity with rate reported in India (37.6%) [18] but lower than rates reported in Nepal (47%) [19], Iraq (64.8%) [20] and other in low-income countries [4]. The disparity in rates of bacteremia among neonatal intensive care unit (NICU) patients is a widespread occurrence that is shaped by a complex interplay of contributing factors that worsen in hospital setting [11, 12, 44]. In Ethiopia, the elevated incidence of neonatal bacteremia can be linked to a range of factors including deficiencies in healthcare infrastructure, economic constraints causing limited access to maternal health services, inadequate infection control measures, spiraling antibiotic resistance, suboptimal diagnostic capabilities and cultural influences on newborn care practices [45]. Bacteremia prevalence disparities are evident globally and within Ethiopia’s regions, with Oromia having statistically significant highest prevalence rate [46]. These variations could be attributed to factors like study design, sample size, detection capacities, population density, malnutrition prevalence, health-seeking behaviors, and the effectiveness of infection prevention and control measures in healthcare facilities [47]. Longitudinal designs are often found to have higher prevalence rates in studies on bacteremia compared to cross-sectional designs. This is because longitudinal designs allow researchers to study the same group of individuals over a longer period, providing a more comprehensive understanding of prevalence and patterns. These designs also typically employ larger sample sizes, which increases statistical power and the likelihood of detecting significant associations or changes. By collecting data at multiple time points, longitudinal designs establish temporal relationships, aiding in the understanding of bacteremia development and progression. They also offer better control over individual differences and confounding factors, reducing their impact on prevalence. In contrast, cross-sectional designs have smaller sample sizes, provide a snapshot of a specific time, and are less effective in capturing changes and establishing temporal relationships. Accordingly, the utilization of a longitudinal study design in Oromia might have contributed to the higher reported prevalence of bacteremia in the region [48]. Pathogenically, Staphylococcus aureus, an individual bacterium belonging to the gram-positive organism, is highly abundant from the current analysis. However, a greater diversity of bacterial species, primarily gram-negative bacteria, is involved, with notable contributions from Klebsiella species and Escherichia coli. This finding is consistent with a study conducted in Nepal, where Staphylococcus aureus accounted for 37.22% of the pathogens identified, followed by Klebsiella pneumoniae at 27.01% and Escherichia coli at 19.70% [16]. However, lower percentages have been reported in other countries such as Pakistan, Nepal, and Iraq, where Klebsiella pneumoniae (34%) and Klebsiella species (22%) were predominantly detected in blood cultures [17, 19, 20]. In Ethiopia, Staphylococcus aureus is a dominant pathogen in neonatal sepsis, likely attributed to poor infection control in NICUs and inadequate essential newborn care. The bacterium was commonly found due to colonization of healthy individuals, vertical transmission during birth, hospital-acquired infections in neonatal units, and the presence of antibiotic-resistant strains like Methicillin-Resistant Staphylococcus aureus [49, 50]. Likewise, several risk factors contribute to the distribution and spread of Klebsiella spp., leading to their dominance in certain regions. Some of the factors are inadequate infection control measures, overcrowding, poor hygiene practices, prolonged hospital stays, invasive procedures, inherent antibiotic resistance and the use of broad-spectrum antibiotic [51]. In prospect of clinical applicability, the significant heterogeneity and variable pathogen distribution observed in this study indicates that simply averaging results into a pooled prevalence rate may not accurately represent true prevalence across neonatal populations [52]. To explore and minimize this heterogeneity, we employed stratified analysis based on study design and region, utilized random effects models to account for variability, and conducted sensitivity analyses to assess the impact of inclusion criteria and analysis methods. Quality assessments of individual studies helped identify potential biases, and publication bias tests, such as funnel plots, ensured no relevant studies were missing. These approaches collectively addressed the heterogeneity issue from multiple angles. Thus, clinicians should be cautious in generalizing results to individual cases, as demographic characteristics, local risk factors, and treatment protocols can vary significantly. This variability emphasizes the need for tailored treatment approaches and highlights the importance of ongoing research to recognize differences in bacteremia prevalence and intervention effects in various contexts, ultimately guiding more personalized clinical decision-making. It is evident that finally, given the current global emphasis on neonatal bacteremia research; it is concerning that there is a dearth of published studies on proven bacteremia in Ethiopia. This deficiency in local research makes it difficult to conduct systematic reviews and meta-analyses, limiting the ability to compare findings and identify broader trends. Consequently, researchers have relied on international prevalence data, which may not accurately represent Ethiopia’s unique healthcare context and patient demographics. This highlights the need for careful consideration of the result when making inferences for neonates in Ethiopia and beyond. Lastly, the systematic review and meta-analysis conducted by the team benefitted significantly from the expertise of seasoned authors, thereby enhancing the potential impact on neonatal survival outcomes by disseminating our results within the scientific community. However, it is essential to acknowledge the limitation of our analysis being confined to observational study designs and lacking representation from all regions of the country. By encompassing data from diverse geographical areas, the robustness and applicability of our findings could be further bolstered. Looking ahead, a more diversified approach incorporating various study designs and broader geographical coverage has the potential to yield nuanced insights that are pertinent to clinical decision-making and public health strategies for this vulnerable demographic. Conclusion In conclusion, the existing studies on proven neonatal bacteremia in Ethiopia were limited in their scope and sample sizes. However, the available evidence clearly indicates that neonatal bacteremia is a significant public health concern in the country, with varying prevalence rates observed across different regions and study designs. Gram-positive bacteria, particularly Staphylococcus aureus and Coagulase-Negative Staphylococcus, have been consistently identified as the primary causative agents while a wider range of gram-negative bacteria, including Klebsiella pneumoniae, Escherichia coli, and Pseudomonas aeruginosa have also been found to be frequently involved in causing the condition with rates varying. Therefore, it is vital to consider the context in diagnosing and treating neonatal bacteremia, especially regarding empirical antibiotic therapy that targets a wide range of pathogens, with a particular focus on gram-negative species. There is also a need for regular surveillance systems, targeted prevention strategies, and effective diagnostic stewardship, along with further research into regional variations and bacterial profiles utilizing various epidemiological study methodologies to effectively enhance the current policy framework and the ominous future. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1