Differences in antibiotic susceptibility among gut-colonizing Acinetobacter baumannii isolated from hospital and community populations in Kenya

¹University of Nairobi, Department of Medical Microbiology and Immunology, Nairobi, Kenya, ²University of Nairobi Institute of Tropical and Infectious Diseases, Nairobi, Kenya, ³Washington State University Global Health-Kenya, Nairobi, Kenya, ⁴Centre for Microbiology Research, Kenya Medical Research Institute, Nairobi, Kenya, ⁵Paul G. Allen School for Global Health, Washington State University, Pullman, WA, USA

^&Corresponding author
Robert Maina Mugoh, University of Nairobi, Department of Medical Microbiology and Immunology, Nairobi, Kenya

Introduction    

The increasing levels of antimicrobial resistance (AMR) pose a serious threat to global public health [1]. Acinetobacter baumannii is one of the ESKAPE pathogens, alongside Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Enterobacter species. In 2017, the World Health Organization (WHO) designated A. baumannii as a critical priority pathogen due to its frequent association with multidrug resistance and its substantial role in healthcare-associated infections [2,3]. The organism´s ability to tolerate a range of temperatures and its capacity to persist on surfaces for extended periods contribute to its endemic presence in healthcare environments, where it can lead to various infections, including bacteremia, meningitis, urinary tract infections, and wound infections [4]. Although A. baumannii primarily occurs in healthcare settings; it is also found in community settings, particularly in tropical and subtropical regions. Individuals with underlying comorbidities such as diabetes, alcoholism, smoking, lung disease, and renal failure are at heightened risk of acquiring these infections [5,6].

Multidrug-resistant strains of A. baumannii exhibit resistance to multiple classes of antibiotics, including aminoglycosides, quinolones, penicillins, cephalosporins, and carbapenems [7]. The rising prevalence of carbapenem-resistant A. baumannii (CRAB) in recent years is concerning, as it severely restricts treatment options; carbapenems are typically reserved as a last-resort treatment for multidrug-resistant infections. The elevated rates of multidrug resistance (MDR) could be largely attributed to the presence of plasmid-encoded β-lactamases, reduced porin numbers and sizes that limit drug entry, and increased expression of efflux pumps [8].

The scarcity of effective drugs for treating A. baumannii infections present challenges for healthcare facilities worldwide [9]. Additionally, many low- and middle-income countries lack the diagnostic capacities necessary to accurately isolate and identify A. baumannii accurately, which hampers the implementation of appropriate treatment protocols, such as antibiotic susceptibility testing (AST) and antimicrobial stewardship programs [9,10].

Colonization with extended-spectrum β-lactamase (ESBL)-producing A. baumannii is often asymptomatic but increases the risk of multidrug-resistant infections and facilitates transmission to non-colonized individuals. Notably, even individuals without prior exposure to healthcare facilities or antibiotics, such as travelers returning from sub-Saharan Africa and South Asia, are colonized with ESBLs, underscoring the significant colonization pressure within communities [5,6,10].

This study aimed to assess the antibiotic susceptibility profiles of A. baumannii isolated from asymptomatic participants enrolled in the Antimicrobial Resistance in Communities and Hospitals (ARCH) study, conducted in selected hospitals and communities in Kenya between January 2019 and March 2020. We also investigated the presence of genes associated with resistance to cephalosporins and carbapenems. Despite the increased risk of colonization with multidrug-resistant organisms, there is limited information on A. baumannii colonization in hospitals and communities globally, particularly in Africa. The scarcity of published information underscores the continued public health relevance of these findings, five years later.

Methods     

Study setting: the study utilized archived stool and rectal samples from the ARCH study conducted between January 2019 and March 2020. The ARCH study enrolled participants from urban (Kibera) and rural (Asembo) settings. Mbagathi County Hospital, which serves Kibera and neighboring communities, was selected for hospital enrollment. In Asembo, inpatients from Siaya County Referral Hospital, Bondo Sub-County Hospital, and St. Elizabeth (Lwak) Hospital were enrolled. Our previous publication provides further insights into the study sites [11].

Sample collection and processing: samples were cultured on HardyCHROM™ Carbapenem Resistant Enterobacterales (CRE) agar plates and HardyCHROM™ Extended-spectrum beta-lactamase (ESBL) agar plates to selectively isolate presumptive A. baumannii. After incubating at 37°C for 18-24 hours, smooth off-white colonies, morphologically consistent with A. baumannii, were subcultured on TSA and further incubated for 18 - 24 hours at 37°C. Purified isolates were emulsified in 50% glycerol phosphate-buffered saline and archived at -80°C for subsequent analysis.

Bacterial identification and AST: archived isolates were revived by thawing at -20°C overnight and at 4°C for 1 hour before processing at room temperature (18-25°C). Isolates were cultured on TSA plates and incubated aerobically at 37°C for 18-24 hours. Bacterial suspensions were prepared from each archived isolate by emulsifying single colonies in 3mL of 0.45% saline and adjusting to a 0.5 McFarland standard using the DensiCHEK Plus (bioMérieux, France). The suspension was then loaded into the VITEK®2 automated system (bioMérieux, France), and isolate identity was confirmed using the Gram-negative ID VITEK®2 automated system cards (bioMérieux, France). Antibiotic susceptibility testing was conducted on A. baumannii isolates using the VITEK®2 AST-GN71 card (bioMérieux, France), which contained a panel of 11 antibiotics (Ampicillin/sulbactam; SAM≤ 4/2 μg/mL, Cefazolin CEZ≤ 4/2 μg/mL, Cefepime FEP ≤ 2 μg/mL, Ceftriaxone CRO≤ 1 μg/mL, Ciprofloxacin CIP≤ 0.5 μg/mL, Gentamicin GEN ≤ 4 μg/mL, Imipenem IMP≤ 1 μg/mL, Meropenem MEM ≤ 0.5 μg/mL, Sulfamethoxazole/trimethoprim SXT ≤ 1 μg/19 μg/mL, Tigecycline TGC ≤ 0.75μg/mL, Tobramycin TOB ≤ 8μg/mL). Minimum inhibitory concentration (MIC) values were interpreted according to the 2020 Clinical Laboratory Standards Institute guidelines [12]. Although A. baumannii is intrinsically resistant to cefazolin due to chromosomal mechanisms; this antibiotic was included in testing to identify any potential extended-spectrum β-lactamases (ESBLs) or other acquired resistance genes that could modify susceptibility.

Detection of resistance genes: K. pneumoniae ATCC 700603 was used as a positive control strain for both phenotypic and genotypic assays. Nuclease-free water was used as a negative control for molecular assays. Deoxyribonucleic acid (DNA) was extracted using the DNeasy Ultra Clean Kit (Qiagen, Maryland, USA), a bead-based method for the isolation of genomic DNA, following the manufacturer´s instructions. Briefly, bacterial cells were suspended in a bead solution-containing tube, and beads were added along with the lysis solution. On vortexing, microorganisms were lysed by a combination of heat, mechanical energy, and detergent. DNA from lysed cells is bonded to a silica spin filter, washed, and recovered in DNA-free Tris buffer. The extracted DNA was stored at -80°C until analyzed using conventional PCR. Table 1 shows the primers used to screen isolates for resistance genes encoding resistance to cephalosporins (bla_CTX-M, bla_SHV, and bla_TEM) and carbapenems (bla_NDM1, bla_OXA-23, and bla_OXA-48) [13] using conventional PCR.

A singleplex assay was conducted using 25 μL reaction mix containing 12 μL of PCR master-mix (Taq DNA polymerase supplied in a proprietary reaction buffer (pH 8.5), dATP 400 μM, dGTP 400 μM, dCTP 400 μM, dTTP 400 μM, MgCl2 23 mM) (Promega, Madison, USA) with pairs of specific primers (10 pmol) at 1μL each, both forward and reverse primers (Table 1), and 1μL DNA template and 12 μL nuclease-free water. PCR was conducted using the Gene AMP 9700 PCR system (Applied Biosystems, USA) in 0.2ml PCR tubes. Amplification was done by initial heating at 95°C for 5 minutes, followed by 30 cycles of denaturation at 95°C for 30 seconds, annealing at the specified primer temperature for 30 seconds, extension at 72°C for 2 minutes, with a final extension of 72°C for 7 minutes. Visualization of amplified PCR products was performed using a 1.5% agarose gel with SYBR Green dye, along with a 1kb molecular ladder. The gel was run for 60 minutes in 5X Tris-acetate Ethylenediaminetetraacetic Acid buffer at 90 volts, 65 mA, and 6 watts. Bands corresponding with respective sequence amplicon sizes (Table 1) were visualized under a UV transilluminator.

Data analysis: data from the VITEK®2 system were exported to MS Excel for cleaning and management. Isolates with resistance to 3^rd or 4^th generation cephalosporins with MIC values ≥ 32 μg/mL for cefepime, ≥ 64 μg/mL for ceftriaxone, and ≥ 32 μg/mL for ceftazidime were presumed to likely harbor ESBLs. Similarly, isolates resistant to carbapenems with MIC ≥ 8 μg/mL for ertapenem, ≥ 8 μg/mL for meropenem, or ≥ 8 μg/mL for imipenem were considered likely to carry carbapenemases. Isolates that were resistant to at least one antibiotic from three or more antibiotic classes were classified as multidrug-resistant [8]. Data were analyzed using SPSS version 26.0 (IBM Corp., NY, USA). Differences in AST results were compared using Fisher´s exact test, and the presence and distribution of genes encoding β-lactam resistance were evaluated across community and hospital isolates.

Ethical considerations: ethical approval was sought from the KNH-UON Ethics and Research Committee (ERC) and approved under Ref no. P504/06/2022. Archived isolates were obtained from stool samples from a study previously approved by KNH-UoN ERC Ref no. P164/03/2018. An ethics review committee waiver of informed consent was obtained.

Results     

A total of 138 presumptive Acinetobacter isolates were selectively revived and microbiologically confirmed, of which 125 viable isolates were identified as A. baumannii. These comprised 86 (69%) community isolates (54 urban and 32 rural) and 39 (31%) hospital isolates (21 urban and 18 rural) (Table 2).

Antimicrobial susceptibility: all A. baumannii isolates, regardless of setting, were resistant to cefazolin but susceptible to tigecycline. Significant differences in antibiotic resistance were observed between community and hospital isolates. Resistance was higher in hospital isolates for ampicillin/sulbactam, ciprofloxacin, ceftriaxone, cefepime, gentamicin, imipenem, and meropenem (all p < 0.001), as well as sulfamethoxazole/trimethoprim (p = 0.003; Table 3). No statistically significant differences in resistance were observed between urban and rural hospital isolates, or urban and rural community isolates (all p > 0.05). Except for cefazolin and sulfamethoxazole/trimethoprim, all isolates from rural community participants were susceptible to all other antibiotics tested. Four (22%) rural hospital isolates, and 8 (38%) urban hospital isolates were classified as multidrug-resistant; no community isolate showed multidrug-resistant phenotypes. Six MDR profiles were identified, with CEZ-CIP-CRO-FEP-GEN-IMP-MEM-SAM-SXT being the only shared profile, observed in 5 (63%) urban isolates and 1 (25%) rural isolate. Other profiles occurred less frequently and were restricted to either urban or rural hospital isolates (Table 4).

Among the 12 phenotypic MDR isolates, seven carried at least one antimicrobial resistance gene detected by PCR, while the remaining five did not show any of the screened resistance genes. Of the six targeted genes associated with beta-lactam and carbapenem resistance (bla_CTX-M, bla_NDM1, bla_OXA-23, bla_OXA-48, bla_SHV and bla_TEM), only two were detected: bla_CTX-M in five urban isolates and bla_NDM1 in two rural isolates.

Discussion     

Antimicrobial resistance remains a major global public health threat, with colonization by drug-resistant bacteria increasing the risk of infection and transmission in healthcare and community settings. This study aimed to investigate and compare the antibiotic susceptibility profiles, resistance patterns, and genetic determinants of gut-colonising A. baumannii isolates from hospital and community settings across urban and rural areas in Kenya. The findings highlight significant disparities in antibiotic resistance between hospital and community A. baumannii isolates, with hospital isolates showing substantially higher resistance rates to multiple clinically important antibiotics such as ampicillin-sulbactam, ciprofloxacin, ceftriaxone, cefepime, gentamicin, imipenem, meropenem, and sulfamethoxazole-trimethoprim. These findings align with observations from other studies that reported high resistance profiles in hospital settings [14,15], suggesting that selective pressure from antibiotic use within hospitals may play a significant role in shaping these resistance patterns.

In contrast, rural community isolates had low resistance levels, exhibiting susceptibility to nearly all antibiotics tested, except for cefazolin and sulfamethoxazole/trimethoprim. Similar observations have been reported in veterinary medicine [16] and may reflect a relatively stable environmental reservoir of A. baumannii in rural settings, where the selective pressure from broad-spectrum antibiotic use is minimal [7]. This "baseline susceptibility" in these communities may provide insights into the evolutionary trajectory of resistance, highlighting the impact of human activity in urban and hospital environments. Furthermore, universal resistance to cefazolin across all isolates, regardless of origin, suggests its intrinsic resistance to this antibiotic, a known characteristic of A. baumannii strains [17]. Conversely, the universal susceptibility to tigecycline highlights its potential as a last-line treatment option against A. baumannii in these settings.

The exclusive presence of multidrug-resistant phenotypes in hospital isolates in this study raises concerns regarding the possibility of diminishing treatment options over time. The identification of six distinct MDR phenotype profiles among hospital isolates, with the CEZ-CIP-CRO-FEP-GEN-IMP-MEM-SAM-SXT profile being the only shared phenotype between urban and rural hospitals, highlights a complex evolution of resistance within these environments. Additionally, since no significant differences in resistance were observed between urban and rural isolates within either the hospital or community settings, we surmise that the immediate environment, specifically high antibiotic use in hospitals versus presumably lower use in the community, exerts a stronger selective pressure than the broader geographical context [11]. These findings indicate that the micro-environment may be the more critical unit of analysis, and that infection control and antibiotic stewardship interventions may need to be tailored at a more specific, hospital-level rather than relying on broader regional strategies, as has been suggested by other studies involving this bacterium [15].

The detection of specific resistance genes, such as bla_CTX-M in five urban isolates and bla_NDM1 in two rural isolates, provides important insights into the molecular mechanisms underlying resistance to beta-lactam and carbapenem antibiotics. In this study, the reference strain used as a positive control amplified for the specific ESBL-associated gene but not carbapenemase genes, consistent with its known genotype. Notably, the absence of other targeted resistance genes in the hospital and community isolates, particularly bla_OXA genes, which have been reported globally as the predominant mechanism of carbapenem resistance in A. baumannii raises questions about the genetic determinants driving resistance in these populations. The observed partial overlap between the phenotypic MDR and genotypic detection of resistance genes reported in this study may be explained by alternative resistance mechanisms, such as efflux pump overexpression, porin loss or hyperproduction of AmpC β-lactamases [18]. This observation aligns with findings from previous studies conducted in Kenya that demonstrated the emergence and persistence of MDR A. baumannii in hospital settings. Studies by Huber et al. [19] and Revathi et al. [20] reported the presence of bla_NDM1 and bla_OXA-23 producing A. baumannii in hospital settings in Kenya, indicating early dissemination of carbapenem-resistant strains in the region. Similarly, investigations by Musyoki et al. [21] and Musila et al. [22] documented widespread resistance across multiple antibiotic classes and circulation of carbapenemase genes, including bla_OXA-23, bla_NDM-1 and bla_OXA-58, among high-risk sequence types such as ST1, ST2, and ST1475.

Limitations: this study had several limitations. First, the sample selection was limited to a few sites, and the isolates selectively represented a resistant subpopulation and not the total colonization burden, which limits the generalizability of the findings. Additionally, the study primarily focused on specific resistance genes, potentially overlooking other mechanisms contributing to antibiotic resistance. Future investigations could incorporate whole-genome sequencing and broader genetic analyses to identify and characterize these mechanisms. Lastly, a longitudinal approach would guide assessment changes in resistance over time, providing insights into the evolving dynamics of A. baumannii resistance.

Conclusion     

This study underscores the need for robust infection control and antibiotic stewardship measures in hospitals to address multidrug-resistant A. baumannii. While isolates from hospitals had high levels of resistance, rural community isolates had near-universal susceptibility, indicating a less affected reservoir. Further, the distinct multidrug-resistant phenotypes in urban and rural hospitals highlight the significance of local factors over geographical distinctions. Our limited identification of resistance genes implies that other non-targeted mechanisms play a role in resistance transmission in these settings. Future strategies should prioritize tailored interventions based on specific local resistance patterns.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Robert Mugoh: conceptualization, methodology, and visualization. Winnie Mutai, Victor Moses Musyoki, Sylvia Omulo: oversight and leadership responsibility for the research activity planning and execution, including mentorship. Charchil Ayodo, Beatrice Oduor, Moureen Jepleting, Susan Kiiru: conducting research and investigation process, specifically performing the laboratory experiments and data collection. Robert Mugoh and Teresa Ita: formal analysis. All the authors have read and agreed to the final manuscript.

Acknowledgments     

We acknowledge Washington State University, whose parent project-the Antibiotic Resistance in Communities and Hospitals (ARCH) study-we acquired the isolates analysed in this study. We also thank the University of Nairobi Institute of Tropical and Infectious Diseases (UNITID) for hosting the project, as well as the Centre for Microbiology Research at the Kenya Medical Research Institute in Nairobi, Kenya, where molecular testing was conducted.

Tables     

Table 1: primers used to detect antimicrobial resistance genes in Acinetobacter baumannii isolates

Table 2: antibiotic susceptibility patterns of Acinetobacter baumannii isolates across the different study sites

Table 3: antibiotic susceptibility among Acinetobacter baumannii isolates from hospital and community sites

Table 4: distribution of multidrug resistance phenotypes among Acinetobacter baumannii hospital isolates

References