Background

Klebsiella species are aerobic or facultative anaerobic gram negative bacteria which belong to the family Enterobacteriaceae. It is known to produce plasmid mediated extended spectrum beta lactamases (ESBLs) which breaks down antibiotics into inactive form [1] and the most common isolated species include Klebsiella pneumoniae, Klebsiella oxytoca, Klebsiella ozaenae and Klebsiella rhinoscleromatis [2, 3]. From the various species, Klebsiella pneumoniae is the most common and ubiquitous in nature causing various human infections [4, 5].

The widespread use of antimicrobials in clinical practice has led to the emergence of resistant bacterial pathogens contributing to the increased morbidity and mortality observed worldwide [6, 7]. Resistant Klebsiella is one of the opportunistic pathogen showing frequent acquisition of resistance to antibiotics accounting to about one-third of all Gram-negative infectious diseases [8, 9] such as bloodstream infection, pneumonia, urinary tract infections (UTI), nosocomial and community acquired infections [2, 10]. As there is inappropriate use of antimicrobials, resistance is tremendously increasing and the therapeutic option has been significantly reduced [11, 12]. Ultimately, this increases cost of treatment and impedes the effective prevention and treatment outcomes in clinical settings [13,14,15,16].

Nowadays, studies show that Klebsiella has been resistant to most common antibiotics including cephalosporins, monobactams, fluoroquinolones and aminoglycosides [3, 17,18,19]. This is because of the frequent empirical use of antibiotics; and persistent exposure of Klebsiella to a number of antimicrobial agents which facilitate the emergence of drug-resistant strains [20]. Previous studies have indicated Klebsiella’s resistance to antibiotics reaching 68.3% in south Africa [12], 54% in India [1], and 97.17% in Equatorial Guinea [21]. In sub-Saharan Africa including Ethiopia, the anti-microbial drug resistance is a serious problem [22]. Individual studies conducted in Ethiopia showed that the prevalence of antimicrobial resistance is high and pooled prevalence of resistance in gram negative bacteria Klebsiella pneumonia is reported to be 23.2% [23]; Klebsiella species are able to develop cross resistance; and the treatment failure is also high [3]. Moreover, one study indicates that antimicrobial resistance level of the gram-negative bacteria ranges from 20 to 100% [24]. Studies in some Ethiopian hospitals such as Jimma (77.8%) [25], Yekatit 12 (64.7%) [26] and Gondar hospitals (95.6%) [27], show that the antimicrobial resistance of Klebsiella is high. However, the comprhenssive analysis of antimicrobial drug resistance pattern of Klebsiella isolates from different parts of Ethiopia has not yet been performed nationally. Hence, the present systematic review and meta analysis was aimed at establishing the pooled prevalence of Klebsiella resistamce; and antimicrobial-specific resistance pattern among Klebsiella clinical isolates in Ethiopia.

Methods

This meta-analysis was carried out in a similar approach to the previously published studies [13, 28] and the content of this systematic review and meta-analysis is well described according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flow diagram and checklist [29,30,31] (S1 file).

Study selection

Exhaustive literature search was done systematically in PubMed, and Science direct, Google scholar; and manual search from Google of potentially relevant articles. The search was done by four authors (LG, TT, GGK and KBT) independently; and the search strategy was built by combining the three main arms (Table 1): Klebsiella, Antimicrobial related terms, and Ethiopia. From the citation extracted, abstracts were scanned to retrieve the clinical studies on Klebsiella infection. Studies that were relevant, by title and abstract, were accessed in full text to determine articles to be included in our meta-analysis. Finally, the references cited by each eligible study were scrutinized to identify additional source; and references were cited using endnote citation manager software (version X7)[Niles software, Clarivate analytic company]. Before starting the data extraction, the protocol was developed and we went directly to data extraction without considering the registration, because at that time there was internet interruption because of the conflict here in Tigray, Ethiopia.

Table 1 Search arms and terms used as a searching strategy
Full size table

Inclusion and exclusion criteria

Studies included in this comprehensive meta-analysis were those that had extractable data on drug resistance of Klebsiella isolates taken from human samples in Ethiopian health facilities or research centers. Articles published from January 2009 to December 2019 in English language; and primary researches with cross-sectional study designs were included; but review papers and other studies that were not relevant to the outcomes of interest and those that were not based on Ethiopian setting were excluded. Initially, non- relevant studies were excluded based on their titles and abstracts. From the screened papers, the duplicates, incomplete data, and studies with very small number of isolates and tested antimicrobials (less than 5 isolates and tested drugs) were excluded.

Outcome of interest

The primary outcome of interest was the prevalence of antimicrobial resistance of Klebsiella species among the total Klebsiella clinical isolates in which the prevalence was calculated by dividing the numbers of resistant Klebsiella isolates by the total number of clinically isolated Klebsiella. We also calculated the pooled resistance pattern of Klebsiella isolates to specific antibiotics as a secondary outcome of interest.

Data extraction

The process of screening (by title, abstract, and full text and data extraction) was done independently by three authors (LG, GGK, and TT), and at each step KBT was involved in consensus creation in cases of discrepancies among the three authors. During screening by title and abstract, the authors reviewed the full text of the article if needed. Data from eligible studies were extracted and summarized into an excel spreadsheet. For each of the included studies, the following information was extracted; name of regions, study design, study area/city, study names, study period, study design, types of specimens, study population, number of study participants, total numbers of isolated Klebsiella species, the average percentage of resistant Klebsiella species, antimicrobial resistance rate of Klebsiella species and references. As all of the articles used in this study are cross-sectional, the score for the quality of the study was assessed using the modified Newcastle-Ottawa Scale (NOS)[32, 33] for the representativeness of sample, appropriateness of sample size, response rate, validity of method, strategy to control confounding factors, reliability of outcome determination, and appropriate statistical analyses. The quality score (Table 2) disagreements were resolved by consensus and a final agreed-upon rating was assigned to each study (S2 file). Moreover, from the included studies each antibacterial tested for Klebsiella was extracted (Table 3).

Table 2 Summary of 35 studies reporting the prevalence and resistance pattern of Klebsiella in different parts of Ethiopia, 2009–2019
Full size table
Table 3 Percentage of pooled antibiotic resistance rates of Klebsiella; Ethiopia, 2009–2019
Full size table

Quality control

The quality of the included studies was checked independently by two authors (LG and KBT) using a set of predetermined criteria such as research design quality of paper, completeness of extractable information, and employed methods for Klebsiella species isolation.

Data analysis

A random-effects analysis method was employed to determine the pooled prevalence, subgroup analysis, and 95% confidence interval (CI) using the approach of Der- Simonian and Laird [34]. Variances and CIs were stabilized using Freeman-Tukey arc-sine methodology [35]. Heterogeneity of study results was assessed using I2 test and significant heterogeneity was considered at p < 0.10 and I2 > 50 [34, 36]. Open Meta-Analyst (version 3.13) and Comprehensive Meta-Analysis (version 3.1) statistical analyses were used. Moreover, subgroup analyses based on administrative regions and mechanism of antibacterial action were also performed to improve the specificity of the assessment of the tested drugs.

Result

We obtained 174 potentially relevant studies through searching electronic databases. From these, 88 duplicate articles were removed by the help of Endnote(version X7)[Niles software, Clarivate analytic company); and the remaining 86 records were screened using their title and abstract out of which 35 articles were omitted based on the exclusion criteria. Full texts of 51 records were evaluated for eligibility in which 16 articles were removed due to data incompleteness, small number of isolates and/or tested antibiotics (less than 5 Klebsiella isolates and antibiotics in a single study). Finally 35 eligible studies were included for the meta-analysis (Fig. 1).

Fig. 1
figure 1

PRISMA flow diagram depicting the selection process for meta-analysis

Full size image

Cross-sectional study design was used in all the included 35 studies. A total of 13,269 study samples were employed, from which 1017 Klebsiella species were isolated. The included studies were taken from Amhara, Oromia. Southern Nations, Nationalities and People (SNNP), Tigray and Addis Ababa; however, there was no study obtained from other regions of Ethiopia (Afar, Benishangul-Gumuz, Gambella, Somali, Harari and Dire Dawa city administration). For screening of Klebsiella species, various specimens were utilized from the various part of the human body, including blood, urine, sputum, body fluids, ear discharge, wound swab, eye swab, stools and pus (Table 2).

The paper-based analysis of this study showed that the overall Klebsiella resistance in Ethiopia was 53.75% (95% CI: 48.35—58.94%) (Fig. 2). Subgroup analyses were carried out based on the region (Addis Ababa, Amhara, Oromia, SNNP, and Tigray), and then the average prevalence of Klebsiella resistance was determined in region wise. Southern Nations, Nationalities, and Peoples of Ethiopia was ranked first (64.39%, 95% CI: 54.94–73.83%), followed by Addis Ababa (55.67%, 95% CI: 47.74–63.40%), Oromia (55.24%, 95% CI: 43.95–66.50%), Amhara (50.1%, 95% CI: 40.82–59.20%), whereas relatively low (but highly varied) prevalence of Klebsiella resistance was reported from Tigray region (46.16%, 95% CI: 21.97–70.34%) (Figs. 3 and 4). The level of heterogeneity of the included studies was high, by random model methods (I2 = 90.55%; P < 0.01).

Fig. 2
figure 2

Forest plot of pooled percentage of Klebsiella antimicrobial resistance in 35 studies, Ethiopia, 2009–2019

Full size image
Fig. 3
figure 3

Proportion of Klebsiella resistance in diferent regions of Ethiopia, 2009–2019. Values in parenthesis indicated 95% CI of Klebsiella resistance in diferent regions of Ethiopia

Full size image
Fig. 4
figure 4

Subgroup analysis of Klebsiella antibacterial resistance according to regions of Ethiopia

Full size image

Additional subgroup analyses were done to estimate the pooled prevalence of Klebsiella resistance based on the mechanism of action of the antibacterial drugs. Antimetabolite (a single drug, trimethoprim-sulfamethoxazole) accounted the highest resistance percentage (66.91%, 95% CI: 59.79–74.06%), followed by cell wall synthesis inhibitors (61.61%, 95% CI: 44.79–79.42%), protein synthesis inhibitors (45.95%, 95% CI: 27.29–64.62%), whereas nucleic acid synthesis inhibitors showed relatively low resistance percentage (35.98%, 95% CI: 31.83–40.14%) and level of heterogeneity of this class was not significant (I2 = 0%, p = 0.68). With regarding to individual antibiotics, high resistance rates were observed to ampicillin (90.56%, 95% CI: 86.31–94.81%), followed by amoxicillin (76.01%, 95% CI: 61.44–90.38%), trimethoprim-sulfamethoxazole (66.91%, 95% CI: 59.76–74.06%), ceftazidime (65.07%, 95% CI: 52.68–77.46%). A relatively low level of resistance rate was observed to amikacin (16.74, 95% CI: 6.84–26.64), and cefoxitin 29.73%, 95% CI: 12.05–47.41%) (Fig. 5).

Fig. 5
figure 5

Subgroup analysis of pooled percentage and confidence interval of Klebsiella resistance to antibacterial drugs according to drug mechanism of action

Full size image

Discussion

The nationwide meta-analysis of Klebsiella resistance was not done in Ethiopia so far. So we conducted a comprehensive systematic review and meta-analysis to determine the pooled of Klebsiella antimicrobial resistance in Ethiopia. From our findings Klebsiella displayed diverse resistance patterns, with percentages varying slightly based on the antibiotic type and geographical distribution. Based on this meta-analysis, the overall Klebsiella resistance in Ethiopia was found to be 53.75% (95% CI: 48.35—58.94%). Similar findings were reported from meta-analysis study done in sub-Saharan Africa and Iran, where the resistance of Klebsiella for some antibiotics was greater than 50% [17, 20]. This indicates more than half of the isolated Klebsiella species were resistance to the most common antibacterial drugs. The possible reasons could be over/under dose of antimicrobial drugs in treating infectious diseases, and empiric use of antibiotics which would inevitably lead to emergence of resistance [24].

Antimicrobial resistance is becoming a public health challenge in the 21st century as a result of many factors including misuse of antimicrobials by health professional and patients, inadequate surveillance systems [7, 27]. Particularly, the emergence and spread of resistant strains of Klebsiella species are a considerable threat to public health [37]. For Klebsiella, resistance to antimicrobials is a normal evolutionary process, but the resistance rate is frequently aggravated by misuse of antibiotics [13, 38]. As per previous study, from the different species of Klebsiella, Klebsiella pneumoniae is becoming known for its resistance properties to most of the last-line antibiotics [8].

In this meta-analysis, the subgroup analysis of regional prevalence of Klebsiella indicated that the highest prevalence of Klebsiella resistance (64.39%) was estimated in SNNP, which is relatively higher than Tigray region (46.16%) (Fig. 4). The observed variation might be due to differences in study location, hospital setup, and antimicrobial utilization. The level of heterogeneity of the included studies was high, by random model methods (I2 = 90.55%; P < 0.01). This indicates that, the included studies have been done in different study areas, study periods, and study populations, which could have an effect on the heterogeneity of the studies.

The other subgroup analysis, based on the mechanism of action of antimicrobials, showed that Klebsiella exhibited higher resistance to antimetabolites (66.91%), which is actually represented by single drug (trimethoprim-sulfamethoxazole), followed by cell wall synthesis inhibitors (61.61%)(Fig. 5). In our finding, there was only one drug (trimethoprim-Sulfamethoxazole) from the antimetabolites tested for Klebsiella resistance, which might have led to the highest value of resistance group compared to the others. However, the highest resistance rate of Klebsiella was noted from individual antibiotics grouped under cell wall synthesis inhibitors including ampicillin (90.56%), amoxicillin (76.01%) and amoxicillin-clavulanic acid (56.92%).

Our finding was concordant with studies done in West Africa where Klebsiella was resistant to Ampicillin (92.5%) and amoxicillin-clavulanic acid (66.1%) [39], and a little bit lower Klebsiella resistance to ampicillin was reported from the three districts of Uganda (66.5%) [40] and India [1]. The higher resistance rate of Klebsiella species to ampicillin and amoxicillin could possibly be due to the fact that the bacteria naturally produces extended spectrum beta-lactamases which inactivate beta lactam` antibiotics (ESBLs) [11, 20]. Our finding showed almost similar pooled Klebsiella resistance to ceftriaxone (47.6%) with previous meta-analysis done on wound infection in which more than half of Klebsiella pneumoniae isolates exhibited resistance to ceftriaxone (57%) [23].

The findings of this study indicated that there is high magnitude of Klebsiella antimicrobial resistance in Ethiopia. The possible reasons could be limited infection surveillance programs, the lack of communication between physicians and microbiologists, lack of standardized criteria to determine drug resistant isolates, limited laboratory facilities, and poor sanitation [20]. The relatively high rates of drug resistant isolates of Klebsiella seen in this meta-analysis might have negative effects on public health which could cause difficulty in treating Klebsiella pneumonia associated infections since only fewer effective drugs are available for treating these highly drug-resistant strains [20]. Hence, functional infection control committee, applying infection prevention protocols, advocating rational prescribing habits, appropriate antimicrobial therapy, health education; and improvement of personal and environmental hygiene need to be applied to curb the resistance problem [13, 18, 20, 41].

Some limitations in our study should be acknowledged. The pooled resistance of Klebsiella was not calculated based on species basis as there was shortage of studies done on individual species. Many of the included studies describe Klebsiella pneumonia, but rarely for other species. Therefore, this meta-analysis should be seen in the context of such limitations.

Conclusion

In this systematic review and meta-analysis, the pooled Klebsiella resistance was found to be considerably high (53.75%) to most of the essential antibiotics in Ethiopia. Klebsiella was highly resistant to ampicillin and amoxicillin but relatively lower to amikacin. Therefore, implementing proper antibiotic prescription policies and appropriate antimicrobial therapy could be the potential interventional strategies to address the emerging resistance of Klebsiella species.