Molecular characterization of virulence and antibiotic resistance genes in Klebsiella pneumoniae isolated from sputum samples at a tertiary hospital in Ethiopia

Abstract

Klebsiella pneumoniae, a Gram-negative, encapsulated rod and prominent member of the ESKAPE pathogen group, ranks among the leading causes of bacterial pneumonia worldwide. It has become increasingly problematic due to multidrug resistance (MDR) and the emergence of hypervirulent strains. In Ethiopia, there are growing reports of the increasing prevalence of MDR K. pneumoniae strains. However, molecular data on the resistance mechanisms and virulence determinants among clinical isolates remain limited. This study molecularly identified K. pneumoniae in sputum samples from patients clinically suspected of pneumonia at a tertiary hospital in Ethiopia, and evaluated their antibiotic resistance profiles, as well as molecularly characterized the isolates for key virulence and antimicrobial resistance (AMR) associated genes. A total of 182 sputum samples from pneumonia- suspected patients were collected and processed following standard microbiological procedures for bacterial isolation and antimicrobial sensitivity testing. K. pneumoniae isolates were confirmed by MALDI-TOF and PCR. Antibiotic susceptibility was assessed via the Kirby–Bauer disk diffusion method. Key virulence and AMR associated genes were detected via PCR. Among the 182 sputum samples, 32 K. pneumoniae isolates were identified, 94% of which were MDR. The predominant resistance genes detected included blaCTX-M (40.6%), blaNDM (34.4%), blaSHV (31.3%), and blaTEM (18.8%). Among the virulence genes, fimH was found in 56.3% of the isolates, mrkA in 28.1%, and rmpA in 9.4%. Additionally, the major outer membrane porin gene ompK35 and the mdtK efflux pump gene were detected in 62.6% and 28.1% of the isolates, respectively. This study reveals a high prevalence of MDR K. pneumoniae strains, and emergence of hypervirulent phenotypes posing a significant threat to therapeutic efficacy. The findings highlight complicated resistance mechanisms driven by molecular synergies, underscoring the urgent need for enhanced molecular surveillance, infection control, and antibiotic stewardship in healthcare settings.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-39069-3.

Introduction

The discovery of antibiotics in the 20th century represented a transformative medical breakthrough, dramatically improving global health outcomes through effective bacterial infection treatment¹. Nonetheless, the emergence of drug-resistant microorganisms poses a significant challenge to these advances, with antimicrobial resistance (AMR) now recognized as a critical global public health threat. Among these pathogens, Klebsiella pneumoniae (K. pneumoniae) stands out as a prominent MDR pathogen^2–4. In 2024, the WHO recognized carbapenem-resistant K. pneumoniae (CRKP) as a critical priority pathogen because of its extensive resistance and significant threat to public health⁵.

K. pneumoniae can colonize healthcare environments and medical devices. This bacterium poses a significant public health challenge because of its increasing pathogenicity, virulence, and emergence of antimicrobial resistance. The association of K. pneumoniae with invasive procedures and its ability to form biofilms increases the risk of infection and complicates treatment options^6,7. The emergence of antibiotic-resistant strains with enhanced virulence poses dual threats to both immunocompetent and immunocompromised individuals^8–11. The increasing number of infections and emergence of MDR strains of this bacterium have attracted increasing attention.

The emergence of MDR K. pneumoniae is driven by selective pressure, widespread antibiotic use, and the conjugational transfer of antibiotic resistance genes across different bacterial species and genera. AMR in K. pneumoniae is complex and involves multiple mechanisms rather than a single mechanism^12,13. Bacteria can acquire resistance through diverse strategies. These mechanisms include the modification of antibiotic target sites, the production of drug-hydrolyzing or drug-inactivating enzymes, the activation of efflux pumps that expel antibiotics, and the prevention of the drug from reaching its target site. These combined mechanisms often act synergistically to increase resistance and render many currently available antibiotics ineffective^14–16.

K. pneumoniae can produce these drug-hydrolyzing or inactivating enzymes by expressing chromosomally encoded genes or genes from plasmid acquisition¹⁷. Horizontal gene transfer significantly contributes to the acquisition of resistance genes by pathogens. The dominant mechanism in this respect is the acquisition of plasmids that confer novel resistance^18,19. The presence of multiple resistance genes on a single plasmid has been documented, contributing to the development of MDR strains²⁰.

K. pneumoniae is a significant global pathogen and an agent of both nosocomial and community-acquired infections. It is strongly associated with a wide range of clinical conditions, including urinary tract infections, pneumonia, septicemia, pyogenic liver abscesses, wound infections, meningitis, endophthalmitis, and lung abscesses. The bacterium contributes substantially to the burden of lower respiratory infections (LRIs), particularly pneumonia. A major challenge in managing these infections is the increasing prevalence of antimicrobial resistance. The emergence and global dissemination of MDR strains, including CRKP, have severely limited treatment options and worsened patient outcomes. Moreover, hypervirulent strains of K. pneumoniae are being identified with increasing frequency, adding further complexity to disease management^{8–11,21–24}.

In Ethiopia, K. pneumoniae is responsible for approximately 22% of bacterial pneumonia cases, reflecting the global increase in MDR and CRKP, particularly in healthcare settings. The pooled prevalence of MDR Klebsiella species in Ethiopia is estimated at 72%, with rates reaching as high as 97% in Addis Ababa. This rising resistance contributes to prolonged hospital stays, increased mortality, and increased healthcare costs²⁵. This high prevalence is alarming, given the frequent reports of multidrug resistance in clinical isolates. This situation mirrors global trends, with increasing detection of MDR and CRKP strains in Ethiopian clinical environments. Although data are limited, there is growing concern about the potential emergence and dissemination of hypervirulent K. pneumoniae strains. These strains can increase disease severity and complicate treatment²⁶.

Despite evidence of the significant role of K. pneumoniae in infection and increasing antibiotic resistance in Ethiopia, few studies have examined the molecular characteristics of these isolates. The genetic basis of antimicrobial resistance in this context remains underexplored, with limited data available on both resistance mechanisms and virulence factors. Understanding these molecular features is vital for elucidating transmission patterns and resistance mechanisms, which in turn are critical for developing targeted interventions and optimizing treatment strategies. To address this gap, the current study focuses on the molecular characterization of K. pneumoniae isolates obtained from sputum samples of pneumonia-suspected patients at Tikur Anbessa Specialized Hospital (TASH), with a particular emphasis on identifying associated antimicrobial resistance genes and virulence determinants.

Materials and methods

Description of the study area

The study was conducted at Tikur Anbessa Specialized Hospital (TASH) in Addis Ababa, Ethiopia, from February 2024 to April 2025. The laboratory work was carried out at the Health Biotechnology Laboratory, Biotechnology Research Centre, Addis Ababa University. Established in 1964, TASH is the largest referral and teaching hospital in the country and is affiliated with Addis Ababa University’s School of Medicine. It provides highly specialized clinical services nationwide. The hospital has approximately 700 to 800 beds and manages a substantial patient volume, with over 500,000 outpatient visits and more than 21,000 inpatient admissions annually. As a major referral center, TASH serves patients from diverse regions across Ethiopia²⁷.

Study design and sample collection

A cross-sectional study was conducted from February 10 to April 20, 2024, to isolate K. pneumoniae from patients suspected of having pneumonia. Sputum samples were collected consecutively from patients suspected of having pneumonia at TASH. Patients suspected of having bacterial pneumonia and who were able to provide sputum samples were eligible for inclusion in this study. Patients who had taken antibiotics within 72 h before sample collection were excluded. Additionally, patients who were unable to provide a sufficient sample volume were excluded from the study. Pneumonia-suspected patients were identified by clinical signs of lower respiratory infection, such as cough, fever, difficulty breathing, chest pain, and abnormal breath sounds, with radiological evidence of lung infiltrates considered when available.

A total of 182 sputum samples were collected. The samples were collected and processed according to standard laboratory protocols²⁸. Patients were provided with clean, sterile, dry, wide-necked sputum containers and were instructed to cough deeply to produce sputum samples. The samples were labeled with a unique code and the date of collection. The appearance of the sputum was categorized as purulent, mucopurulent, mucoid, mucosalivary, blood-tinged, and currant jelly-like, as described by Cheesbrough et al.²⁸. Data were collected via the Kobo Toolbox mobile application.

The samples were immediately transported to the laboratory in an ice box. Approximately 1 mL of the sputum sample was inoculated directly into 7 mL of double-strength tryptone soy broth (TSB; HiMedia, India). The mixture was subsequently incubated at 37 °C for 18–24 h to increase bacterial recovery^29,30. The following day, a loopful from the enrichment mixture was inoculated onto a MacConkey agar plate via a sterile loop. After the plate was incubated at 37 °C for 24 h, non-duplicate presumptive colonies were selected for further identification. Large, mucoid, lactose-fermenting colonies were specifically chosen. The presumed colonies were then inoculated onto 5% sheep blood agar plates (BAPs) and Eosin-Methylene Blue (EMB) (HIMEDIA, India) and incubated at 37 °C for 24 h. Colonies exhibiting pink, dark-centered, and mucoid characteristics on EMB agar, as well as large, mucoid, nonhemolytic colonies on blood agar (BAP) plates, were selected for Gram staining and microscopic examination. The isolates were further confirmed via standard biochemical tests, including urease, indole, motility, and citrate tests. The characteristics of the colonies on all media were recorded. The isolates were stored at -20 °C in TSB containing 25% glycerol (HIMEDIA, India) for further characterization. Biochemical tests, including urease, citrate utilization, indole, and motility tests, were performed to characterize the isolates according to the methods of Osman et al.³¹. Cellular morphology and Gram reactions were assessed via Gram staining and examination via oil immersion microscopy. The isolates were further confirmed as K. pneumoniae by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry and polymerase chain reaction (PCR).

Isolate identification by MALDI-TOF MS

Isolates presumed to be K. pneumoniae by biochemical tests were subjected to identification via matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS). For sample preparation, a direct extraction method utilizing formic acid was employed. A fresh colony grown on 5% sheep blood agar was picked with a disposable microbial loop and transferred to a spot on the MALDI 96 target (Zybio EXS3000, Inc., China). The colony was smeared to form a thin layer, overlaid with 1 µL of 70% formic acid, and allowed to air dry at room temperature. Once dried, the spots were covered with 1 µl of α-cyano-4-hydroxycinnamic acid (α-CHCA) matrix dissolved in a standard solvent (SS) composed of 47.5% HPLC-grade water, 2.5% trifluoroacetic acid, and 50% acetonitrile to achieve a final concentration of 10 mg/ml. For calibration, E. coli ATCC 8739 was used, with the strain incubated for 18–24 h at 35 ± 2 °C on blood agar under standard conditions. According to the manufacturer’s instructions, log scores of 2.0 or higher indicate reliable species-level identification. Log scores below 2.0 but equal to or greater than 1.7 indicate genus-level identification or presumptive species-level identification. Log scores < 1.7 are considered unreliable. Isolates that had log scores less than 1.7 were rechecked by reculturing.

Molecular detection and characterization of K. pneumoniae

Genomic DNA extraction and quality assessment

Genomic DNA was extracted from 32 MALDI-TOF-confirmed K. pneumoniae isolates via the boiling method with minor modifications (proteinase K treatment), as described by Belete et al.³². Briefly, the isolates were cultured on tryptone soy agar (TSA) (HIMEDIA, India) for 24 h at 37 °C. Two to three colonies were inoculated into an Eppendorf tube containing 100 µL of 1× TE buffer (10 mM Tris-HCl, 1 mM EDTA). The mixture was thoroughly mixed to dissolve the colonies and then centrifuged at 13,000 rpm for five minutes at 4 °C. The supernatant was discarded, and 100 µL of nuclease-free water was added to the pellet. The mixture was then boiled at 95 °C for 10 min in a water bath. The preparation was immediately incubated at -20 °C for 30 min³². Then, 3 µL of proteinase K was added, and the mixture was gently mixed at room temperature. After two minutes of incubation at room temperature, the preparation was centrifuged at 13,000 rpm for five minutes at 4 °C. Fifty microliters of the resulting supernatant, which contained the DNA, was carefully transferred to a new Eppendorf tube and used directly as the template DNA for PCR amplification. The quantity and quality of the extracted DNA were assessed via a Nanodrop 2000/2000 C Spectrophotometer (Thermo Scientific™, USA) and 0.8% agarose gel electrophoresis, respectively. The extracted DNA was stored at -20 °C for future molecular analysis.

Molecular characterization of K. pneumoniae

Molecular identification of K. pneumoniae, along with detection of virulence factors and AMR genes, was performed using PCR. For identification, a specific primer pair targeting the 16–23 S intragenic transcribed spacer (ITS) region of K. pneumoniae was used³³. AMR and virulence genes were also assessed via the primers listed in Table 1. All amplifications were performed in an A300 Thermal Cycler (LongGene Scientific Instruments, China). The specific cycling protocols for each primer are described in Supplementary file 1. The PCR products were analyzed via electrophoresis on a 1.5% (w/v) agarose gel (LA agarose, Jena Bioscience, Germany) and stained with ethidium bromide solution (10 mg/ml) (HIMEDIA Laboratories, India). Electrophoresis was conducted at 100 V for 1 h in 1X TAE buffer (40 mM Tris, 20 mM acetic acid, and 1 mM EDTA at pH 8.0). Ten microliters of the PCR product was mixed with 6X loading dye. A 100 bp DNA Ladder (Thermo Scientific™ GeneRuler™, Waltham, Massachusetts, USA) was used to estimate the size of the PCR products. The resulting band patterns were visualized and documented via a UVITEC Cambridge gel documentation system (Cambridge, United Kingdom).

Table 1.

List of primers used for molecular characterization of K. pneumoniae.

S. no	Target	Sequence 5’ − 3’	Product length (bp)	Ref.
1	16 S rRNA	FW: ATTTGAAGAGGTTGCAAACGAT RV: TTCACTCTGAAGTTTTCTTGTGTTC	130	34
2	gyrA	FW: CGCGTACTATACGCCATGAACGTA RV: ACCGTTGATCACTTCGGTCAGG	441	35
3	RmpA	FW: ACTGGGCTACCTCTGCTTCA RV: CTTGCATGAGCCATCTTTCA	535	36
4	blaCTX-M	FW: TTTGCGAGTGCAGTACCAGTAA RV: CTCCCCTGCCGGTTTTATC	519	37
5	OmpK35	FW: CTCCAGCTCTAACCGTAGCG RV: GGTCTGTACGTAGCCGATGG	241	36
6	blaSHV	FW: TTAACTCCCTGTTAGCCA RV: GATTTGCTGATTTCGCCC	796	37
7	blaTEM	FW: ATGAGTATTCAACATTTCCG RV: CTGACAGTTACCAATGCTTA	867	38
8	blaKPC	FW: CGTCTAGTTCTGCTGTCTTG RV: CTTGTCATCCTTGTTAGGCG	798	39
9	blaNDM	FW: GGTTTGGCGATCTGGTTTTC RV: CGGAATGGCTCATCACGATC	621
10	ArmA	FW: CCGAAATGACAGTTCCTATC RV: GAAAATGAGTGCCTTGGAGG	846	40
11	Mdtk	FW: GCGCTTAACTTCAGCTCA RV: GATGATAAATCCACACCAGAA	543	41
12	fimH	FW: CGATCACTGACTACGTCACC RV: CCGTGAATCGTAAACCACC	143	42
13	mrkD	FW: GCCTTAATGCTGATGCCATTAC RV: AACCACTGACACTGACTCCC	180
14	mrkA	FW: CACCAAACAGGATGATGTGAG RV: CGCATAGCCGACGTAGTAAG	262
15	Mcr-1	FW: AGTCCGTTTGTTCTTGTGGC RV: AGATCCTTGGTCTCGGCTTG	320	43

FW: forward, RV: reverse, bp: base pair.

K. pneumoniae antimicrobial susceptibility (AST) profiling

Antimicrobial susceptibility testing (AST) was conducted using AST discs (Oxoid™, United Kingdom). A Mueller‒Hinton agar plate was used to perform disc diffusion testing according to the 33rd edition of the CLSI guidelines⁴⁴. The AST test was conducted on 15 available antibiotics, including cefotaxime (CTX) (30 µg), kanamycin (K) (30 µg), aztreonam (ATM) (30 µg), tetracycline (TE) (30 µg), meropenem (MEM) (10 µg), piperacillin/tazobactam (TZP) (100/10 µg), ertapenem (ETP) (10 µg), ciprofloxacin (CIP) (5 µg), sulfamethoxazole/trimethoprim (STX) (1.25/23.75 µg), cefoxitin (FOX) (30 µg), doripenem (DOR) (10 µg), nitrofurantoin (FM) (100 µg), ceftazidime (CAZ) (30 µg), gentamycin (GEN) (30 µg), and amikacin (AN) (30 µg). To conduct AST, newly cultured confirmed K. pneumoniae colonies incubated at 37 °C for 24 h were suspended in sterile normal saline to obtain a homogeneous suspension equivalent to 0.5. The suspension was adjusted to the McFarland standard. The mixture was subsequently spread over the entire surface of Mueller‒Hinton agar via a sterile cotton swab. Afterward, the antibiotic discs were placed at regular intervals on a 90 mm petri dish. After the plates were incubated for 18 h at 37 °C, the diameter of the growth inhibition was measured via a standard caliper (TM-52000-86). The diameters of the antibiotic susceptibility zones were interpreted according to the 33rd edition of the CLSI guidelines⁴⁴, and the results were reported as resistant (R), intermediate (I), or sensitive (S). K. pneumoniae isolates resistant to one antibiotic from three or more different classes of antibiotics were identified as MDR, and the isolates that were resistant to nearly all the tested antibiotic classes were classified as extensively drug resistant (XDR), as previously described ⁴⁵⁴⁵⁴⁵. Escherichia coli ATCC^® 25,922 and K. pneumoniae ATCC^® 700,603 were used as quality control strains. The control strains were collected from the Ethiopian Public Health Institute. The multiple antibiotic resistance (MAR) index was calculated and interpreted according to Reza et al. ³¹³¹³¹ via the following formula: a/b, where “a” represents the number of antibiotics to which the isolate was resistant and “b” represents the total number of antibiotics tested.

Phenotypic confirmation of ESβL-producing K. pneumoniae isolates

K. pneumoniae isolates resistant to third-generation cephalosporins were screened for further confirmation of extended-spectrum β-lactamase (ESβL) production. For ESβL confirmation, a combined disc test (CDT) was conducted according to the 33rd edition of the CLSI guidelines⁴⁴. Two to three colonies of newly cultured K. pneumoniae, obtained after incubation at 37 °C for 24 h, were dissolved in 3 mL of sterile normal saline to obtain a homogeneous suspension comparable to the McFarland 0.5 standard. The suspension was subsequently spread on the whole surface of Mueller‒Hinton agar. Antibiotic discs containing cefotaxime (CTX) (30 µg), ceftazidime (CAZ) (30 µg), cefotaxime + clavulanic acid (CTX/CLA) (30 µg/10 µg), and ceftazidime + clavulanic acid (CAZ/CLA) (30 µg/10 µg) were subsequently placed on the prepared lawn at a 30 mm distance from their corresponding combination. After the plates were incubated for 18 h at 37 °C, the diameter of the bacterial growth inhibition around the cephalosporin and its clavulanic acid combination was measured via a standard caliper (TM-52000-86), and the difference was calculated. An ESβL producer was confirmed to have an inhibitory effect on zones with a diameter difference > 5 mm⁴⁵.

Phenotypic confirmation of carbapenemase-producing K. pneumoniae

The detection of carbapenemase enzyme production was conducted via the modified carbapenemase inactivation method (mCIM) and the EDTA-modified carbapenemase inactivation method (eCIM) as previously described⁴⁶. Carbapenemase detection was performed on all carbapenem-resistant and intermediate K. pneumoniae isolates during the AST test. Meropenem antibiotic disks (Oxoid™, United Kingdom) were used to confirm carbapenem production. Briefly, for the mCIM test, a 1-µL aliquot from a fresh K. pneumoniae isolate was transferred to 2 mL of tryptic soy broth (TSB) via a sterile inoculation loop and suspended. A meropenem (MEM) disc (10 µg) (Oxoid™, United Kingdom) was then dipped into the bacterial suspension. The mixture was incubated at 35 °C for 4 h, with a tolerance of ± 15 min. Following incubation, the disc was aseptically removed via a sterile inoculation loop. It was then positioned onto a Mueller‒Hinton agar plate that had been previously inoculated with a lawn of susceptible E. coli (ATCC 29522). The results were recorded after 18–24 h of incubation at 37 °C. For mCIM-positive K. pneumoniae isolates, the eCIM test was conducted to differentiate between serine-β-lactamase and metallo-β-lactamase (MβL) enzymes. Twenty microliters of 0.5 M EDTA (HiMedia, India) was added to one tube, and the other tube was kept EDTA-free. A 1-µL aliquot from the mCIM-positive strain was transferred into each tube via a sterile loop. The meropenem (MEM) disc (10 µg) (Oxoid™, United Kingdom) was then immersed in the bacterial suspension and incubated at 35 °C for 4 h ± 15 min. After incubation, the disc was aseptically removed with a sterile inoculation loop and placed onto a Mueller‒Hinton agar plate seeded with a lawn of susceptible E. coli (ATCC 29522). The results were observed and recorded after 18–24 h of incubation at 37 °C. The eCIM result was interpreted only for the mCIM result, which indicates the presence of carbapenemase. A zone of inhibition of 6 to 15 mm around the disk or a zone of inhibition of 16 to 18 mm that also contained colonies was considered positive. Conversely, a zone of inhibition measuring 19 mm or greater was considered negative. This finding indicates that resistance to carbapenems may be due to a mechanism other than carbapenemase production. If the mCIM showed an increase in diameter of ≥ 5 mm compared with the mCIM, the isolate was classified as mCIM positive⁴⁴.

Screening for the hypermucoviscous (hpV) K. pneumoniae phenotype

Screening for the hypermucoviscous phenotype was performed as previously described⁴⁷. Briefly, the isolates were cultured on a blood agar plate. After incubation at 37 °C for 24 h, a sterile inoculation loop was used to contact a colony from the agar plate. The mucoid colony was gently stretched by pulling the loop upward. A positive string test was indicated by the formation of viscous strings greater than 5 mm in length between the loop and the colony. No string formation or strings less than 5 mm are considered negative for the hypermucoviscous phenotype.

Ethical consideration

The study was conducted in accordance with all applicable institutional, national, and international ethical standards. Ethical approval was obtained from the Research Ethics Committee of the Institute of Biotechnology, Addis Ababa University (Reference: IoB/431/2016/2024). Permission for conducting data collection was duly authorized by the Chief Executive Officer of Tikur Anbessa Specialized Hospital. All procedures involving human participants adhered to the principles of the Declaration of Helsinki. Written informed consent was obtained from each participant after a thorough explanation of the study’s aims, potential risks, and benefits. Participant confidentiality was strictly maintained by coding all the data and samples, which were used exclusively for research purposes. The participants were informed that the risks involved were minimal and that the study did not require any special tests or procedures, causing emotional distress beyond routine medical care.

Statistical analysis and data visualization

Statistical analysis and data visualization were performed via the R programming language (version 4.5.1; r-project.org) in RStudio. A p value less than 0.05 was considered statistically significant. Analyses were conducted via base R functions and packages such as dplyr (for data manipulation) and ggplot2 (for visualization).

Results

Isolation and identification

A total of 182 sputum samples were collected from pneumonia-suspected patients at Tikur Anbessa Specialized Hospital (TASH). Among the recruited patients, 100 (55%) were male, and 82 (45%) were female. Initially, 42 isolates were presumed to be K. pneumoniae by biochemical tests. However, after confirmation via MALDI‒TOF MS and PCR, the actual recovery of K. pneumoniae was 17.6%. Although the occurrence of K. pneumonia was greater among male patients (10.44%) than among female patients (7.14%), the difference did not reach statistical significance (p = 0.67). Of the 41 presumptive K. pneumoniae isolates identified based on preliminary biochemical and morphological characteristics, 32 were confirmed as K. pneumoniae by MALDI-TOF MS and PCR.

Among the different types of sputum samples analyzed, the highest percentage recovery of K. pneumoniae was observed in sputum samples with a currant jelly appearance. Out of the 22 currant jelly sputum samples analyzed, 13 tested positive for K. pneumoniae, indicating the highest proportion in comparison to other types (p = 0.00517). In contrast, the lowest percentage recovery was from muco-salivary sputum samples (Fig. 1).

Fig. 1.

Occurrence of K. pneumoniae by sputum type among pneumonia patients at TASH, 2024.

Hypermucoviscous phenotype screening

Among the 32 isolates tested for the hypermucoviscous phenotype, 4 (12.5%) exhibited a positive string test (Table 3). A positive result was defined as the formation of a mucoviscous string exceeding 5 mm in length when a bacterial colony was stretched with an inoculation loop (Fig. 2).

Table 3.

Summary of the molecular and phenotypic profiles of K. pneumoniae isolates.

Isolates (n = 32)	blaCTX-M	blaNDM	blaSHV	blaTEM	blaKPC	Mcr-1	FimH	MrkA	MrkD	Mdtk	gyrA	ArmA	RmpA	OmpK35	CDT - test	mCIM-test	emCIM-test	String test	Phenotype	MAR index
TASP04	+	+	-	-	-	-	+	+	-	+	+	-	-	-	-	+	-	-	MDR	0.86
TASP06	+	+	+	+	-	-	-	+	-	+	+	+	-	-	-	+	+	-	XDR	1
TASP09	+	-	+	-	-	-	-	+	+	+	-	-	+	+	+	+	-	+	MDR	0.733
TASP17	-	-	-	+	-	-	-	-	+	-	-	-	-	-	-	+	-	+	MDR	0.6
TASP18	-	+	+	+	-	-	-	+	-	-	+	+	-	-	+	+	-	-	MDR	0.8
TASP22	-	-	-	-	-	-	+	-	-	-	-	-	-	-	-	-	-	-	MDR	0.466
TASP28	-	-	-	-	-	-	+	-	-	-	-	-	-	-	-	-	-	-	MDR	0.466
TASP32	+	+	+	-	-	-	-	+	+	+	-	-	-	+	-	+	+	-	XDR	1
TASP37	+	+	+	-	-	-	-	-	-	-	+	-	+	+	-	+	+	-	MDR	0.733
TASP41	-	-	-	-	-	-	-	-	-	-	+	-	+	-	-	-	-	+	MDR	0.733
TASP45	-	-	-	-	-	-	+	-	-	-	-	-	-	+	-	-	-	-	MDR	0.4
TASP50	+	-	+	-	-	-	+	-	-	-	-	-	-	-	-	-	-	-	MDR	0.533
TASP54	-	-	-	-	-	-	-	-	-	-	+	+	-	+	-	+	-	-	MDR	0.866
TASP59	-	-	-	-	-	-	-	-	-	-	-	+	-	+	-	-	-	-	MDR	0.6
TASP65	-	-	-	-	-	-	+	-	-	-	+	-	-	+	-	+	-	-	MDR	0.733
TASP72	+	+	-	-	-	-	+	+	-	-	+	+	-	+	+	-	-	-	MDR	0.8
TASP76	+	+	-	-	-	-	+	+	+	+	-	-	-	-	-	+	+	-	MDR	0.66
TASP88	-	-	-	-	-	-	+	-	-	-	+	-	-	+	-	-	-	+	MDR	0.8
TASP92	+	+	+	-	-	-	+	-	-	-	-	-	-	-	-	+	-	-	MDR	0.6
TASP96	-	-	-	-	-	-	+	-	-	-	-	-	-	+	-	-	-	-	MDR	0.26
TASP101	+	+	+	+	-	-	-	+	-	+	-	+	-	+	-	+	+	-	MDR	0.733
TASP109	+	+	+	+	-	-	-	+	-	+	-	+	-	-	-	+	+	-	MDR	0.8
TASP117	+	+	+	-	-	-	-	-	-	+	+	-	-	+	-	+	-	-	XDR	1
TASP121	-	-	-	-	-	-	+	-	-	-	-	-	-	+	-	-	-	-	MDR	0.6
TASP126	-	-	-	-	-	-	+	-	-	+	-	+	-	+	-	-	-	-	MDR	0.733
TASP129	-	-	-	-	-	-	+	-	-	-	-	-	-	-	-	-	-	-	MDR	0.6
TASP 164	+	-	-	+	-	-	+	-	-	-	-	-	-	+	+	-	-	-	MDR	0.4
TASP142	-	-	-	-	-	-	+	-	-	-	-	-	-	+	+	-	-	-	DR	0.066
TASP153	-	-	-	-	-	-	-	-	-	-	-	-	-	+	+	-	-	-	DR	0.133
TASP158	-	-	-	-	-	-	-	-	-	-	-	-	-	+	-	-	-	-	MDR	0.4
TASP174	-	-	-	-	-	-	+	-	-	-	-	-		+	-	-	-	-	MDR	0.66
TASP168	-	-	-	-	-	-	+	-	-	-	-	-	-	+	-	-	-	-	MDR	0.26
Total	13 (40.6%)	11 (34.4%)	10 (31.3%)	6 (18.8%)	0	0	18 (56.3%)	9 (28.1%)	4 (12.5%)	9 (28.1%)	9 (28.1%)	8 (25%)	3 (9.4%)	20 (62.6%)	6(18.8%)	14 (43.75%)	6 (18.8%)	4 (12.5%)

“+” indicates the presence or positive result for the corresponding test, and “-” indicates the absence or negative result for the corresponding test.

Fig. 2.

Typical string-positive hypermucoviscous phenotype.

Antibiotic susceptibility of K. pneumoniae isolates

The isolates exhibited diverse resistance and susceptibility patterns to various antibiotics, as summarized in Table 2. High resistance rates were noted for tetracycline (87.5%) and cefotaxime (84.4%). In contrast, high susceptibility to gentamycin (59.4%) was observed.

Table 2.

Antimicrobial susceptibility testing results for K. pneumoniae isolates (n = 32).

S.no	Antibiotic	Antibiotic susceptibility
		Intermediate (I), No (%)	Resistant (R), No (%)	Susceptible(S), No (%)
1	Meropenem	8 (25)	19 (59.4)	5 (15.6)
2	Piperacillin/Tazobactam (TZP)	5 (15.6)	26 (81.25)	1 (3.1)
3	Amikacin (AN)	12 (37.5)	12 (37.5)	8 (25)
4	Aztreonam (ATM)	6 (18.7)	23 (71.9)	3 (9.4)
5	Kanamycin (K)	11 (34.4)	13 (40.6)	8 (25)
6	Ceftazidime (CAZ)	7 (21.9)	21 (65.6)	4 (12.5)
7	Ertapenem (ETP)	4 (12.5)	17 (53.1)	11 (34.4)
8	Gentamycin (GEN)	3 (9.4)	10 (31.2)	19 (59.37)
9	Tetracycline (TE)	3 (9.4)	28 (87.5)	1 (3.1)
10	Nitrofurantoin (FM)	7 (21.8)	14 (43.8)	11(34.4)
11	Cefotaxime (CTX)	4 (12.5)	27 (84.37)	1 (3.1)
12	Ciprofloxacin (CIP)	6 (18.7)	24 (75)	2 (6.3)
13	Sulfamethoxazole/trimethoprim (STX)	3 (9.4)	24 (75)	5 (15.6)
14	Doripenem (DOR)	7 (21.8)	20 (59.3)	6 (18.7)
15	Cefoxitin (FOX)	6 (18.7)	22 (68.5)	4 (12.5)

Confirmation of ESβL-producing K. pneumoniae

Among the 28 isolates that were resistant and had intermediate susceptibility to cefotaxime (CTX) and ceftazidime (CAZ), only 6 (21.4%) were confirmed to produce ESβLs. Carbapenemase production was performed on isolates that were resistant to at least one of the tested carbapenem antibiotics during the classical antibiotic susceptibility test. Among the 20 carbapenem-resistant isolates assessed for carbapenemase production, 14 tested positive for carbapenemase production via mCIM (Table 3). Among the isolates that tested positive by mCIM, 6 were confirmed as MβL producers through the eCIM test, indicating the presence of metal-dependent carbapenemase activity. According to the eCIM test, 42.86% (6/14) of the isolates were identified as MβL-positive carbapenemase producers, whereas the remaining 57.14% (8/14) may harbor serine carbapenemases or dual mechanisms.

Molecular detection of virulence and AMR genes in K. pneumoniae

Among the 32 isolates tested, blaCTX-M was the most prevalent resistance determinant, detected in 40.6% of the isolates (Table 3). This gene was followed by blaNDM, which was identified in 34.4% of the isolates. blaSHV and blaTEM were present in 31.3% and 18.8% of the isolates, respectively (Annex 6). Neither the blaKPC nor the Mcr-1 gene was detected in any of the isolates. Among the assessed virulence factors, the adhesin gene FimH was the most prevalent and was found in 56.3% of the isolates. Other fimbrial genes, MrkA and MrkD, were detected in 28.1% and 12.5% of the isolates, respectively (Table 3). Genes associated with multidrug efflux pumps (Mdtk), quinolone resistance (gyrA), and aminoglycoside resistance methylase (ArmA) were found in approximately 25% of the isolates. The outer membrane porin gene OmpK35 was detected in 62.6% of the isolates (Table 3). A clustered heatmap of AMR and virulence gene distributions across isolates is shown in Fig. 3.

Fig. 3.

Clustered heatmap showing the distribution of AMR and virulence genes across 32 K. pneumoniae isolates among pneumonia-suspected patients at TASH, 2024.

Discussion

In the present study, 32 K. pneumoniae strains were isolated from the sputum of patients suspected of having pneumonia. The occurrence of K. pneumoniae in our study was 17.6% (32/182) among sputum samples, which falls within the reported range of isolation rates across various regions in Ethiopia²⁵. This rate is somewhat moderate compared with other studies: it is higher than the 4.7% reported in Arba Minch⁴⁸ but lower than the notably high rates reported in Addis Ababa (39.5%)⁴⁹ and Gondar (31.0%)⁵⁰. We used molecular methods to confirm K. pneumoniae isolates, thereby improving accuracy by reducing presumptive biochemical identifications from 41 to 32 confirmed cases. This molecular confirmation likely contributed to the observed differences in isolation rates compared to studies relying solely on biochemical tests, by minimizing false positives and enhancing diagnostic precision. Our findings closely align with the reported overall occurrence of K. pneumoniae in Ethiopia, which is 22%²⁵. The detection rate of hypermucoviscous phenotype (hmv) K. pneumoniae isolates in this study was 12.5% (4/32). In the present study, the highest K. pneumoniae recovery was observed from currant jelly sputum (p = 0.00517), which is consistent with the classic presentation of K. pneumoniae pneumonia characterized by thick, currant jelly sputum resulting from extensive tissue necrosis and inflammation⁵¹. This classical appearance has been reported in many cases of K. pneumoniae infection⁵².

In the present study, 94% of the total K. pneumoniae isolates were MDR, showing resistance to at least three antibiotic classes, and three isolates were classified as extended-spectrum drug resistant (XDR)⁴⁵. Rates of MDR K. pneumoniae ranging from 55.67% to 99% have also been reported in other studies^53–55; however, these studies did not focus exclusively on sputum isolates. In the current study, the high MDR rate likely reflects the tertiary hospital’s role as a referral center, where prior antibiotic exposure and limited diagnostics at primary facilities create selective pressure. Studies have also reported that hospital-acquired infections and AMR gene transfer within healthcare settings significantly contribute to resistance^1,56. The resistance profile observed in this study aligns with recent local and regional research and represents the increasing AMR problem encountered in Ethiopian tertiary care settings⁵⁷.

Resistance to extended-spectrum beta-lactams was greater for cefotaxime (84.4%) and ceftazidime (78.1%). Carbapenem resistance was also high for meropenem (59.4%), ertapenem (53.1%), and doripenem (62.4%), which exceeds previously reported rates of 30–40% and recent data from TASH⁵⁸. This suggests the rapid dissemination of carbapenemase-producing strains, which hydrolyze carbapenems and other beta-lactams, often coupled with other resistance mechanisms^59,60. Aminoglycoside resistance was lower for amikacin (37.5%) but higher for gentamycin (43.8%) and kanamycin (40.6%), which aligns with earlier findings⁵⁴. These findings collectively underscore a consistent pattern in which amikacin has substantially lower resistance rates than do gentamicin and kanamycin, highlighting its potential as a more effective therapeutic option⁶¹. Earlier studies from TASH reported 93.2% susceptibility to amikacin; this difference can be explained by the dynamic nature of antimicrobial resistance, which increases over time owing to selective pressure from antibiotic use⁵⁷. This differential resistance profile underscores the importance of selecting aminoglycosides on the basis of local susceptibility patterns. In this study, the β-lactamase genes blaCTX-M (40.6%), blaSHV (31.3%), and blaTEM (18.8%) were detected, with high blaCTX-M and blaSHV rates, indicating their role in regional ESβL spread. These findings confirm the link between ESβL production and elevated K. pneumoniae resistance, which is consistent with prior research⁶². Notably, a prior study conducted on bloodstream infection isolates at TASH reported that blaCTX-M was present in 67.3% of ESβL-producing Enterobacteriaceae, with K. pneumoniae accounting for only 10% of these cases⁶³. Differences in sample source and a likely increase in blaCTX-M prevalence over time help explain the variation.

Carbapenems are the most potent β-lactams, exhibiting broad activity against both Gram-positive and Gram-negative bacteria. However, the emergence of CRKP has become a global concern⁶⁴. In the present study, blaNDM was detected in 34.4% of all the K. pneumoniae isolates and accounted for 78.5% of the carbapenemase-producing isolates. The blaKPC genes were not detected among the 32 tested K. pneumoniae isolates using a single primer set targeting the predominant KPC-2/3 variants. However, the reliance on this single primer pair constitutes a study limitation, as it may have overlooked rare KPC alleles. Future studies should use multiplex PCR and whole-genome sequencing for comprehensive carbapenemase profiling. Earlier studies at TASH reported blaNDM in 92.9% of carbapenemase producers, with blaKPC detected in only one isolate⁶⁵. This is likely due to the widespread dissemination of blaNDM via mobile plasmids across diverse strains, whereas blaKPC, which is typically linked to specific clones, remains rare in the region.

Among the mCIM-positive isolates, blaNDM was detected in 78.5% of the samples; the remaining samples likely harbored other carbapenemase genes. eCIM identified 42.86% as metallo-β-lactamase (MβL) producers, with 56.2% likely expressing serine-based carbapenemases or both types simultaneously—a pattern supported by studies showing blaNDM coexistence with blaOXA-48, blaKPC, and blaVIM^65–69. Notably, the co-occurrence of these genes was common in our study: blaCTX-M and blaNDM co-existed in 31.3% (10/32), blaCTX-M with blaSHV in 28.1% (9/32), and blaNDM with blaSHV in 25% (8/32). Triple and quadruple gene combinations were also observed, with 9.4% (3/32) harboring all four tested genes (blaCTX-M + blaNDM + blaSHV + blaTEM-1). The co-occurrence of multiple β-lactamase genes, such as those found in K. pneumoniae isolates, can result from their localization either on a single large conjugative plasmid or distribution across multiple plasmids within the same bacterial host²⁶.

Similarly, a previous study from TASH reported a high prevalence of blaCTX-M and frequent co-carriage with blaSHV and blaTEM. Although blaNDM is less commonly documented, it reflects emerging but limited spread⁷⁰. In contrast, studies from other African countries have shown variable blaNDM prevalence but have consistently reported blaCTX-M and blaSHV co-occurrence, which aligns with our findings^71–73. Globally, the co-occurrence of blaNDM with ESβL genes such as blaCTX-M and blaSHV is increasingly reported, which is consistent with our study’s MDR patterns. The complex K. pneumoniae resistome complicates treatment and control^26,65,74. Notably, our study’s triple/quadruple gene combinations are unique, potentially reflecting novel local plasmid architectures that underscore MDR diversity and the need for region-specific molecular surveillance to inform stewardship. The detection of the armA gene in 25% of K. pneumoniae isolates aligns with its role in high-level aminoglycoside resistance (e.g., amikacin, gentamicin) via 16 S rRNA methylation, a mechanism widely reported globally^75,76. Similarly, gyrA mutations were found in 28.1% of the isolates. The gyrA mutation is an established marker of ciprofloxacin resistance via the alteration of DNA gyrA, a commonly known mechanism⁷⁷. To the best of our knowledge, this study provides the first report from Ethiopia on the molecular detection of armA and gyrA in K. pneumoniae isolates.

Antibiotic efflux pumps are among the most important mechanisms of AMR in K. pneumoniae clinical isolates^78–81. The Mdtk complexes are among the best-characterized efflux pumps in K. pneumoniae⁸². In this study, the Mdtk gene was detected in 28.1% of K. pneumoniae isolates, whereas a previous study reported an 88% prevalence of the Mdtk gene in K. pneumoniae strains⁸³. The ompK35 porin gene was present in 62.5% of the total isolates, which is comparable with an earlier study that detected 60% ompK35 in MDR K. pneumoniae⁸³. Notably, all isolates in the present study that harbored Mdtk and/or lacked ompK35 exhibited a multiple antibiotic resistance (MAR) index greater than 0.7, indicating a significant correlation between these genes and high-level antibiotic resistance. This aligns with global findings, where MdtK, a member of the MATE (multidrug and toxic compound extrusion) family, contributes to antibiotic resistance by actively expelling drugs such as quinolones and tetracyclines. Additionally, the loss of ompK35 reduces membrane permeability, limiting antibiotic entry and thereby enhancing resistance, particularly to β-lactams such as cephalosporins and carbapenems^82,84–88. The hypermucoviscous (HMV) phenotype was detected in 12.5% (4/32) of the K. pneumoniae isolates and was characterized by viscous colony strings > 5 mm. In association with increased virulence (e.g., pneumonia, bacteremia, and liver abscesses) via genes such as rmpA and magA, HMV strains occur globally in 20–35% of clinical isolates^47,89,90. This phenotype reflects increased capsule production and enhanced virulence and has been widely reported globally in clinical and community isolates^91–97.

In the present study, the rmpA gene was detected exclusively in three string-positive HMV isolates and was not detected in non-HMV strains. These findings suggest that these non-HMV isolates may lack rmpA-carrying virulence plasmids^{8–11,21–24}. Although the hypermucoviscosity (HMV) phenotype occurs in Ethiopian K. pneumoniae from adult pneumonia, no published data on rmpA or related genes in local strains exist²⁶. Most molecular investigations into hypervirulent K. pneumoniae strains harboring rmpA and other virulence determinants have been conducted in regions such as the Asia-Pacific region and selected African countries^91–99. For example, a study from South Africa reported a rmpA detection rate of 12.2%⁹⁴, which is comparable to the 9.4% reported in our study, although our limited sample size constrains direct comparison. Another study in Kenya also identified virulence factors associated with hypervirulence, further supporting the global distribution of these traits⁹⁵.

K. pneumoniae virulence is increased by the presence of fimbriae^24,100. Fimbria facilitate adherence to surfaces and host tissues. The well-characterized fimbriae in K. pneumoniae are types 1 and 3, which serve as structural and adhesive molecules. Type 1 and type 3 fimbriae have been identified as significant pathogenicity factors during infections⁸. 90% of clinical K. pneumoniae isolates express type 1 fimbriae, which bind D-mannosylated glycoproteins. In our study, the FimH gene was detected in 56.25% of the isolates. In this study, the occurrence rates of mrkA and mrkD were 28.1% and 12.5%, respectively. Earlier studies reported detection rates of 96% for the mrkD gene and 88% for the FimH gene⁸³.

The MrkD adhesin of K. pneumoniae binds to extracellular matrix components such as type IV and V collagen, which are abundant in pulmonary epithelial basement membranes¹⁰¹. This interaction enables K. pneumoniae to attach firmly to the respiratory epithelium, facilitating colonization and biofilm formation in the lung environment^102,103. These data highlight the persistent prevalence of adhesins such as FimH and MrkD, which play critical roles in K. pneumoniae respiratory colonization, biofilm formation, and pathogenicity. Although this single-center tertiary hospital study identified key AMR genes and virulence factors in Ethiopian K. pneumoniae, its generalizability to community settings is limited. Reliance on sputum from pneumonia patients also risks selection bias, as colonization cannot be distinguished from infection without quantitative cultures; multicenter studies with diverse samples are needed.

Conclusion

This study reports the detection of multidrug-resistant and hypervirulent K. pneumoniae isolates from sputum samples of pneumonia-suspected patients in a tertiary hospital in Ethiopia. This study also highlights the widespread occurrence of key resistance genes, including blaNDM and multiple β-lactamases, which contribute to the rapid dissemination of carbapenem resistance. The coexistence of virulence and resistance factors in these strains poses a significant challenge to clinical management and infection control. These findings emphasize the urgent need for continuous molecular surveillance, stringent infection prevention measures, and targeted antibiotic stewardship to control the spread of these high-risk clones in healthcare settings. This work adds valuable molecular data to the existing knowledge on the molecular epidemiology of K. pneumoniae in Ethiopia.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

AAA performed the conceptualization, sample collection, laboratory work, data curation, analysis, and visualization. AGB contributed to the validation, supervision, and critical review of the manuscript. TST was involved in conceptualization, validation, resource provision, supervision, and manuscript reviewing. All authors reviewed and approved the final manuscript.

Data availability

The datasets generated and/or analyzed during the current study are included in this published article and its supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.