Occurrence of antimicrobial-resistant, extended-spectrum beta-lactamase and carbapenemase-producing bacterial isolates from intensive care unit environment in two hospitals of Ethiopia

Abstract

Background

Intensive care unit environments are a critical reservoir of antimicrobial-resistant bacterial species and facilitate the transmission of nosocomial infections. This unit harbors multidrug-resistant organisms, thereby complicating the management of infectious diseases within clinical settings, particularly in resource-limited countries. This study, therefore, aimed to determine the occurrence and antimicrobial resistance profiles of bacteria, including Extended-spectrum beta-lactamase- and Carbapenemase-producing isolates, in the intensive care units of two hospitals in Addis Ababa, Ethiopia.

Methods

A repeated cross-sectional study was conducted in Addis Ababa, Ethiopia, from July 2024 to April 2025. A total of 239 swab samples were collected from adult and neonatal intensive care unit environments during both the dry and rainy seasons. Samples were collected and processed aseptically. Bacterial identification and antimicrobial susceptibility testing were performed according to CLSI guidelines.

Results

From the total samples collected, 162/239 (67.8%) bacterial isolates were recovered from the Intensive Care Unit environment. Klebsiella pneumoniae 22/162 (13.6%) and Pseudomonas aeruginosa 21/162 (13%) were the most frequently isolated bacterial species. All P. aeruginosa were resistant to ceftazidime, while 19/22 (86.3%) and 18/22 (81.8%) of K. pneumoniae were resistant to cefuroxime and cefotaxime, respectively. All S.haemolyticus were also resistant to penicillin, and 8/12 (66.7%) of Enterococcus spp were resistant to erythromycin. Multidrug-resistant bacteria were seen in 89/150 (59.3%) of isolates. The highest multidrug resistance rate was obtained for Pseudomonas aeruginosa 18/21 (85.7%), followed by Klebsiella pneumoniae 16/22 (72.7%) and Acinetobacter spp 8/11 (72.7%). Extended-spectrum β-lactamase-producing isolates accounted for 23/105 (21.9%), and Carbapenemase-producers comprised 10/105 (9.5%).

Conclusion

The environments within intensive care units are conducive to the proliferation of multidrug-resistant microorganisms, encompassing bacterial isolates that produce extended-spectrum beta-lactamases and carbapenemases, thereby complicating the management of infectious diseases. Consequently, it is crucial to address the issue of antimicrobial resistance in this critical setting to mitigate the incidence and dissemination of resistant bacteria, which may lead to the emergence of nosocomial infections.

Introduction

Intensive care units (ICUs) play a critical role in hospital care by treating patients with severe illnesses that require complex monitoring and invasive interventions. Nevertheless, this environment is particularly susceptible to microbial contamination due to invasive procedures, frequent patient handling, the presence of immunocompromised patients, extensive use of medical equipment, and frequent use of broad-spectrum antimicrobials [1]. Pathogenic bacteria can be found on inanimate surfaces and medical equipment in the ICU, raising the possibility of transmission to patients and healthcare personnel. The major factor in the persistence and the spread of nosocomial pathogens is the contamination of high-touch surfaces like bed rails, infusion pumps, ventilators, and catheters [2, 3].

Bacterial colonization in ICU settings originates from different sources. Patients with active infections or asymptomatic colonization release bacteria into the environment. Healthcare personnel unintentionally transfer these bacteria between patients and surfaces through contact with contaminated hands or equipment [4]. Besides, air flow patterns and moisture on surfaces contribute to the survival of bacteria. Common etiological agents involved in ICU-related contamination include many bacterial species such as Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, Enterobacter spp., and E. coli. These bacterial isolates are categorized as ESKAPE-E pathogens and are often responsible for ventilator and catheter-associated infections [5].

Antimicrobial resistance (AMR) is a critical problem facing global public health. Recognizing this problem, the World Health Organization (WHO) has declared AMR as one of the top ten global health challenges facing humanity. It was estimated to cause 4.95 million deaths in 2019, and projections suggest that this number could exceed 10 million by 2050 [6]. AMR, particularly among bacterial isolates in the ICU, poses a significant challenge for infection management. The misuse of antimicrobials in these settings promotes the emergence of drug-resistant strains, such as extended-spectrum beta-lactamases (ESBLs), carbapenemase-producing (CP) bacteria, and methicillin-resistant S. aureus (MRSA), which are major obstacles to treating patients with AMR in clinical settings [7, 8]. Infections caused by these resistant strains are often linked to higher rates of morbidity and mortality, longer hospital stays, and increased healthcare costs.

In sub-Saharan countries like Ethiopia, the scarcity of infection prevention and control (IPC) resources, absence of diagnostic capabilities, and lack of organized, comprehensive data on the burden of AMR may aggravate the crisis of AMR, complicating both empirical treatment and infection control measures [9]. Despite increasing awareness of the role of ICU environmental reservoirs in spreading resistant bacterial species, there is still limited data on the AMR profiles of bacterial isolates from ICU surfaces and devices in Ethiopia. Some studies have documented bacterial contamination from hospital environments in Ethiopia. However, comprehensive data on the burden and magnitude of AMR profiles of bacteria, including ESKAPE-E pathogens, ESBLs, and CP isolates from ICU environments, remains limited, and there is no prior evidence describing the extent and the nature of environmental contamination of ICU environments in the study hospitals.

This study, therefore aimed to determine the occurrence and antimicrobial resistance profiles of bacteria, including the magnitude of ESKAPE-E pathogens, ESBLs, and CP isolates from the ICU environment at Menelik II Comprehensive Specialized Hospital and Tirunesh Beijing General Hospital, Addis Ababa, Ethiopia. By exploring AMR profiles of environmental isolates, this study provides important evidence to inform targeted IPC strategies and promote rational antimicrobial use in ICU settings.

Materials and methods

Study settings

This study was conducted in Addis Ababa, Ethiopia, a city with an estimated population of around six million. Menelik II Comprehensive Specialized Hospital (MIICSH) and Tirunesh Beijing Hospital (TBGH) were purposively selected for their geographical features, capacity to provide a range of medical services to the capital and nearby rural populations, higher bed capacity, absence of existing data, and substantial patient volumes.

Menelik II Comprehensive Specialized Hospital, one of the oldest public hospitals in Ethiopia, is situated in Yeka Sub-city in northeastern Addis Ababa. The hospital has a capacity of 135 surgical beds and serves an estimated catchment population of approximately 1.6 million people. It includes various wards and departments, such as medical, surgical, and intensive care units [10].

Tirunesh Beijing Hospital, established in 2012, is located in Akaki Kality Sub-city and provides healthcare services to residents of Akaki-Kality, neighboring sub-cities, and surrounding communities in the Oromia Regional State. Its primary services include outpatient and inpatient care, emergency services, and intensive care units [11].

Sample collection and preparation

From July 2024 to April 2025, a hospital-based repeated cross-sectional study was conducted during the dry and rainy seasons. Swab samples were collected from both the adult and neonatal ICUs. A total of 239 swab samples were collected from selected high-touch inanimate surfaces and medical devices, including patient beds (bed rails, head and foot boards), linens, suction machines, catheters, ventilators, baby incubators, infusion stands, and surrounding environmental surfaces such as floors and walls adjacent to patients. Additional sites comprised workstations with monitors and keyboards, as well as lobby areas with tables, door handles, light switches, and sinks. Samples were collected through purposive sampling. Out of 239 samples, 120 samples were collected during the dry season, and 119 samples were collected in the rainy season.

All swab samples were collected after the ICU environments had been cleaned and before health care personnel began their activities. Sterile swabs moistened with 85% physiologic saline were used to sample high-touch surfaces following careful screening. Swabbing was conducted in parallel stripes with slight rotation, followed by perpendicular stripes, in accordance with ISO/DIS 14698-1 guidelines [12]. Each sample was then transferred to brain heart infusion broth (BHI) for transport.

Isolation and identification of bacterial isolates

Each swab sample was pre-enriched in BHI for 24 h. Following this, a loopful of turbid broth was cultured onto blood agar (Blood Agar Base, HiMedia, India, with 5% sheep blood) using an inoculating loop, and then incubated for 24–48 h at 37 °C. After proper incubation, each bacterial colony was selected based on its colony morphology, color, and hemolytic properties, and then purified through successive subculturing [13]. Following the acquisition of pure cultures on blood agar (BA), bacterial isolates were categorized as Gram-positive bacteria (GPB) and Gram-negative bacteria (GNB) using Gram staining. GPB were identified at the species level using biochemical tests, such as Oxidation and Fermentation (OF), Oxidase, Catalase, Coagulase, Bacitracin, Optochin, and Salt tolerance test. GNB were cultured on MacConkey (MAC) agar and were subsequently categorized as lactose fermenters (LF) or non-fermenters (LNF) based on their capacity for lactose fermentation as well as their oxidative and fermentative characteristics, which were assessed through the oxidation and fermentation (OF) test. Following this, GNB isolates were identified at the species level utilizing standard biochemical assays, including Oxidase, Catalase, Urease, Indole, Methyl red, Voges–Proskauer, Citrate utilization, Mannitol, Malonate, growth at temperatures ranging from 42 °C to 44 °C, and Lysine iron agar tests, as previously delineated [14, 15].

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) was employed for confirmation and for bacterial isolates that couldn’t be reliably identified using conventional biochemical methods [16]. Finally, the bacterial isolates were preserved in tryptone soy broth supplemented with 30% glycerol and stored at − 80 °C for future analyses.

Phenotypic antimicrobial susceptibility testing

Antimicrobial Susceptibility testing (AST) was performed via the Kirby–Bauer disk diffusion method in accordance with the Clinical and Laboratory Standards Institute (CLSI) guidelines, 34^thedition [17]. A total of 150 bacterial isolates were tested, comprising 45 GPB and 105 GNB. Bacillus cereus among GPB and Aeromonas spp., and Stenotrophomonas maltophila among GNB isolates were excluded due to the absence of interpretive criteria for the zone diameter in these guidelines. GPB isolates were tested against tetracycline (30 µg), gentamicin (10 µg), clindamycin (2 µg), vancomycin (30 µg), cefoxitin (30 µg), penicillin (10 µg), ciprofloxacin (5 µg), erythromycin (15 µg), and sulfamethoxazole–trimethoprim (1.25/23.75 µg). GNB isolates were tested against cefuroxime (30 µg), cefotaxime (30 µg), ceftriaxone (30 µg), ceftazidime (30 µg), cefepime (30 µg), ciprofloxacin (5 µg), sulfamethoxazole–trimethoprim (25 µg), gentamicin (10 µg), amikacin (10 µg), meropenem (10 µg), ertapenem (10 µg), tetracycline (30 µg), piperacillin (100 µg), amoxicillin–clavulanic acid (20/10 µg), aztreonam (30 µg) and piperacillin–tazobactam (100/10 µg). The selection of these antimicrobial agents was based on their availability in the local market and their frequent use in clinical practice. Selection and interpretation followed the CLSI guidelines to ensure consistency and standardization. Multidrug-resistant (MDR) isolates were defined as bacterial isolates exhibiting resistant or intermediate phenotypes to at least one antibiotic agent in three or more drug classes [18].

Phenotypic detection of ESBL-Producing isolates

GNB isolates, particularly enterobacterales that were resistant to at least one third-generation cephalosporin, were screened for ESBL production. Enterobacterales suspected of producing ESBLs exhibited an inhibition zone of ≤ 27 mm for cefotaxime (30 µg) and/or ≤ 22 mm for ceftazidime (30 µg). These isolates were then selected for confirmatory testing using a combination disc-diffusion test (CDT) according to the CLSI guidelines. CDT was conducted using ceftazidime (30 µg), with ceftazidime-clavulanic acid (CAZ/CLA) and cefotaxime (30 µg) with cefotaxime-clavulanic acid (CTX/CLA), following the CLSI guidelines. A difference of ≥ 5 mm in zone of inhibition between CAZ/CLA and CTX/CLA indicated that the organism was an ESBL producer, in accordance with CLSI guidelines [17].

Phenotypic detection of carbapenemase-producing isolates

Enterobacterales showing non-susceptibility to either meropenem or ertapenem and non-fermenter GNB isolates, particularly P.aeruginosa, which exhibited non-susceptibility to meropenem, were screened for carbapenemase production using the modified carbapenem inactivation (mCIM) according to CLSI guidelines [17]. In short, a loopful of bacterial growth from an overnight culture was suspended in 2mL of tryptone soy broth (HiMedia, India). A 10 µg meropenem disk was then added to the suspension and thoroughly immersed. Then the mixture was incubated for 4 h without shaking. After proper incubation, the meropenem disk was removed using an inoculating loop and placed onto MHA (Oxoid, UK) that had been inoculated with a 0.5 McFarland suspension of E.coli ATCC 25,922. Zone diameter of 6–15 millimeters (mm) or the presence of pinpoint colonies within a 16–18 mm zone was used to confirm for CP isolates after overnight incubation in accordance with the CLSI guidelines.

Phenotypic detection of methicillin-resistant S. aureus isolates

The cefoxitin disk diffusion test was employed to identify MRSA isolates. The S. aureus isolates were cultured on MHA using 30 µg cefoxitin disks and a 0.5 McFarland standard suspension. Each plate was incubated overnight at 37 °C. The results were interpreted in accordance with the CLSI guidelines.

Data quality assurance

Quality control measures were applied throughout the entire laboratory procedure to ensure accuracy and reliability of the study findings. The sterility of each batch of newly prepared culture media was evaluated before use. The performance of biochemical tests for bacterial identification was verified using reference strains, which include K. pneumoniae ATCC 700,603, E. coli ATCC 25,922, and S. aureus ATCC 25,923. Strains of E. coli ATCC 25,922 and ATCC 35,218 were used to assess the quality and effectiveness of antimicrobial disks.

AST was performed using control strains such as E. coli ATCC 25922 (ESBL-negative) and K. pneumoniae ATCC 700603 (ESBL-positive) for confirmatory testing of ESBL-producing isolates. In addition, the control strains K. pneumoniae ATCC BAA-1705 (positive) and E. coli ATCC 25,922 (negative) were used to confirm CP isolates. The reference strains were obtained from the Ethiopian Public Health Institute (EPHI).

Data analysis

Data obtained from this study were entered into Microsoft Excel. Statistical analysis and data visualization were conducted via the R programming language (version 4.5.2) in RStudio. Analyses were performed via base R functions and packages such as dplyr and ggplot2 for data manipulation and visualization, respectively. The chi-square test was used to assess the significant difference between the dry and rainy seasons. A p-value less than 0.05 was used for statistical significance.

Results

From the total 239 samples collected in both the dry and rainy seasons, 130/239 (54.4%) were culture positive for one or more bacterial species. From these 130 culture-positive samples, 162 bacterial isolates were recovered. During the dry season, 66/130 (50.8%) of samples were culture-positive. In comparison, 64/130 (49.2%) were also culture-positive during the rainy season. The recovery rate was slightly higher in the dry season, although there wasn’t a statistically significant difference (p-value = 0.850). A higher proportion of bacteria was recovered from TBGH, 70/130 (53.8%), than from MIICSH, 60/130 (46.2%) (Table 1).

Table 1.

Distribution, number of positive and negative samples from intensive care unit environments at Addis Ababa, Ethiopia, 2024/2025

		Total samples (%)	Positive samples (%)	Negative samples (%)
Sampling Hospitals	TBGH	136(56.9)	70(53.8)	66(60.5)
	MIICSH	103(43.1)	60(46.2)	43(39.5)
	Total	239(100)	130(54.4)	109(45.6)
Season	Dry	120(50.2)	66(50.8)	54(49.5)
	Rainy	119(49.8)	64(49.2)	55(50.5)
	Total	239(100)	130(54.4)	109(45.6)

Of the 162 positive bacterial isolates, 52/162 (32.1%) were GPB, and 110/162 (67.9%) were GNB. Among GPB, the dominant isolates were S. aureus 14/52 (26.9%), and the least prevalent GPB were coagulase-negative Staphylococcus spp. 5/32 (3.1%). The dominant isolates among GNB were K. pneumoniae 22/110 (20%), while the least frequently isolated were S. maltophila 2/110 (1.8%). From the total bacterial isolates obtained, 111/162 (68.5%) were ESKAPE-E pathogens. Among the ESKAPE-E group, 85/110 (77.3%) were GNB, while 26/52 (50%) were GPB. From these pathogens, K. pneumoniae 22/85 (25.9%) and S. aureus 14/26 (53.8%) were the dominant among GNB and GPB, respectively. Generally, a total of 15 different bacterial species were identified from ICU environments. K. pneumoniae 22/162 (13.6%) was the most dominant or frequently isolated bacterial species, followed by P. aeruginosa 21/162 (13%) (Table 2).

Table 2.

Frequencies of bacterial isolates recovered from the intensive care unit environments at Addis Ababa, Ethiopia, 2024/2025

GNB Isolates	N (%)	ESKAPE-E Pathogens
A. baumannii	18(11.1)	18(21.2)
Acinetobacter spp	11(6.8)	-
Aeromonas spp	3(1.8)	-
E.cloacae	9(5.6)	9(10.6)
E.coli	15(9.3)	15(17.6)
K.oxytoca	9(5.6)	-
K.pneumoniae	22(13.6)	22(25.9)
P.aeruginosa	21(13)	21(24.8)
S. maltophilia	2(1.2)	-
Total (N)	110(67.9)	85(77.3)

GPB isolates
B.cereus	7(4.3)	-
CoNS	5(3.1)	-
Enterococcus spp	12(7.4)	12(46.2)
S. aureus	14(8.6)	14(53.8)
S.epidermidis	6(3.7)	-
S.haemolyticus	8(4.9)	-
Total (N)	52(32.1)	26(50)

CoNS coagulase-negative Staphylococcus

Magnitude of bacterial pathogens over different inanimate surfaces and devices

The highest bacterial magnitude was found on the patient bed, 23/162 (14.2%), followed by the suction machine, 20/162 (12.3%). The patient bed was predominantly contaminated by Staphylococcus epidermidis, 4/23 (17.4%), followed by K. pneumoniae and CoNS, both at 3/23 (13%). The suction machines were also predominantly contaminated with P. aeruginosa 6/20 (30%), followed by K. pneumoniae 4/20 (20%). The lowest contaminated items were catheters, 3/162 (1.9%). Catheters were colonized with A. baumannii, E. coli, and K. pneumoniae with a similar proportion of 1/3 (33.3%) (Table 3).

Table 3.

Magnitude of bacterial isolates over different animate surfaces and medical devices in Intensive Care Unit Environments at Addis Ababa, Ethiopia, 2024/2025

Bacterial isolates	ICU inanimate surfaces and medical devices (%)
	Baby incubator	Patient bed	Catheter	Door handle	Floor	Infusion pump	Light switch	Linens	Sinks	Suction machine	Ventilator	Wall	Work station
A. baumannii	1(8.3)	1(4.3)	0(0)	2(25)	5(26.3)	0(0)	0(0)	4(23.5)	1(6.7)	3(15)	0(0)	0(0)	1(7.1)
Acinetobacter spp	0(0)	1(4.3)	1(33.3)	1(12.5)	0(0)	1(20)	0(0)	2(11.8)	1(6.7)	1(5)	1(25)	1(6.7)	1(7.1)
Aeromonas spp	0(0)	1(4.3)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	2(10)	0(0)	0(0)	0(0)
B.cereus	2(16.7))	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	3(20)	0(0)	0(0)	2(13.3)	0(0)
CoNS	0(0)	3(13)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	1(16.7)	1(7.1)
E.cloacae	1(8.3)	2(8.7)	0(0)	2(25)	0(0)	0(0)	0(0)	1(5.8)	1(6.7)	0(0)	0(0)	1(16.7)	1(7.1)
E.coli	0(0)	0(0)	1(33.3)	0(0)	3(15.8)	1(20)	0(0)	2(11.8)	1(6.7)	2(10)	1(25)	1(16.7)	3(21.4)
Enterococcus spp	1(8.3))	1(4.3)	0(0)	0(0)	2(10.5)	0(0)	4(57.1)	0(0)	4(26.7)	0(0)	0(0)	0(0)	0(0)
K.oxytoca	1(8.3)	0(0)	0(0)	0(0)	1(5.3)	0(0)	1(14.3)	2(11.8)	0(0)	0(0)	0(0)	1(16.7)	3(21.4)
K.pneumoniae	3(25)	3(13)	1(33.3)	1(12.5)	1(5.3)	1(20)	1(14.3)	3(17.6)	1(6.7)	4(20)	0(0)	0(0)	3(21.4)
P.aeruginosa	1(8.3)	2(8.7)	0(0)	0(0)	2(10.5)	1(20)	0(0)	2(11.8)	3(20)	6(30)	2(50)	2(13.3)	0(0)
S. aureus	2(16.7)	2(8.7)	0(0)	1(12.5)	3(15.8)	0(0)	0(0)	0(0)	0(0)	2(10)	0(0)	4(26.7)	0(0)
S. maltophilia	0(0)	1(4.3)	0(0)	0(0)	1(5.3)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)
S.epidermidis	0(0)	4(17.4)	0(0)	1(12.5)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	0(0)	1(16.7)	0(0)
S.haemolyticus	0(0)	2(8.7)	0(0)	0(0)	1(5.3)	1(20)	1(14.3)	1(5.8)	0(0)	0(0)	0(0)	1(16.7)	1(7.1)
Total	12(7.4)	23(14.2)	3(1.8)	8(4.9)	19(11.7)	5(3.1)	7(4.3)	17(10.5)	15(9.3)	20(12.3)	4(2.5)	15(9.3)	14(8.6)

Antimicrobial resistance profiles of gram-positive bacteria

GPB isolates from ICU environments showed distinct AMR profiles. All S.haemolyticus were resistant to penicillin, and 8/12 (66.7%) of Enterococcus spp were resistant to erythromycin, tetracycline, and penicillin. Of 14 S. aureus isolates, 6/14 (42.9%) were confirmed as MRSA, while 4/12 (33.3%) of the Enterococcus spp exhibited Vancomycin resistance (VRE) (Table 4).

Table 4.

Antimicrobial resistance profiles of Gram-Positive bacteria from intensive care unit environments, Addis Ababa, Ethiopia 2024/2025

GPB (%)	Antibiotics tested
	PEN	AM	ERY	TET	CIP	CLO	SXT	GEN	VAN	FOX
Enterococcus spp(12)	8(66.7)	4(33.3)	8(66.7)	8(66.7)	3(25)	NT	NT	NT	4(33.3)	NT
S.aurues(14)	10(71.4)	NT	5(35.7)	5(45.7)	4(28.6)	2(14.3)	5(45.7)	3(21.4)	NT	6(42.9)
S.haemolyticus(8)	8(100)	NT	2(25)	4(50)	2(25)	1(12.5)	3(37.5)	2(25)	NT	2(25)
S.epidermidis(6)	4(66.7)	NT	2(33.3)	3(50)	1(16.7)	1(16.7)	2(33.3)	1(16.7)	NT	0(0)
CoNS(5)	2(40)	NT	1(20)	0(0)	1(20)	0(0)	2(40)	0(0)	NT	1(20)

Abbreviations: PEN Penicillin, AMP Ampicillin, ERY Erythromycin, TET Tetracycline, CIP Ciprofloxacin, CLO Clindamycin, GEN Gentamycin, VAN Vancomycin, FOX Cefoxitin, NT not tested

Antimicrobial resistance profiles of gram-negative bacteria

GNB isolates exhibited different AMR profiles. All P. aeruginosa were resistant to ceftazidime, while 19/22 (86.3%) and 18/22 (81.8%) of K. pneumoniae were resistant to cefuroxime and cefotaxime, respectively. However, there was no resistance found in K. oxytoca against amikacin, meropenem, aztreonam, and ertapenem 0/9 (0%) (Table 5).

Table 5.

Antimicrobial resistance profiles of GNB from intensive care unit environments at Addis Ababa, Ethiopia, 2024/2025

GNB (%)	Antimicrobials tested
	PIP	AMC/PTZ	FEP	CTX	CRO	CXM	CAZ	AT	ETP	MEM	GM	AN	CIP	SXT	TET
A.baumannii(18)	13(72.2)	10(55.6)	7(38.9)	13(72.2)	13(72.2)	NT	17(94.4)	NT	NT	8(44.4)	8(44.4)	4(22.2)	8(44.4)	9(50)	12(66.7)5)
Acinetobacter spp(11)	6(54.5)	4(36.4)	3(27.3)	6(54.5)	6(54.5)	NT	10(90.9)	NT	NT	2(8.2)	4(36.4)	3(27.3)	4(36.4)	5(45.4)	7(63.6)
E.coli(15)	NT	3(20)	12(80)	10(66.7)	11(73.3)	11(73.4)	12(80)	9(60)	2(13.3)	7(46.7)	10(66.7)	6(40)	4(26.7)	6(40)	11(73.3)
E.cloacae(9)	NT	NT	3(33.3)	2(22.2)	6(66.7)	3(33.3)	1(11.1)	NT	2(22.2)	2(22.2)	2(22.2)	0(0)	1(11.1)	4(44.4)	5(55.6)
K.oxytoca(9)	NT	4(44.4)	1(11.1)	5(55.5)	4(44.4)	5(55.5)	5(55.6)	0(0)	0(0)	0(0)	1(11.1)	0(0)	1(11.1)	3(33.3)	5(55.6)
K.pneumoniae(22)	NT	16(72.7)	11(50)	18(81.8)	15(68.2)	19(86.3)	14(63.6)	6(27.3)	9(40.9)	9(40.9)	11(50)	3(13.6)	6(27.3)	12(54.5)	17(77.3)
P.aeruginosa(21)	16(76.2)	10(47.6)	8(38.1)	19(90.5)	NT	NT	21(100)	3(14.3)	NT	10(47.6)	14(66.7)	5(23.8)	7(33.3)	11(52.4)	13(61.9)

N.B. Piperacillin and Piperacillin/tazobactam were tested only against A. baumannii, Acinetobacter spp., and P. aeruginosa

Abbreviations: PIP Piperacillin, PTZ Piperacillin/tazobactam, AMC Amoxicillin/clavulanic acid, FEP Cefepime, CTX Cefotaxime, CRO Ceftriaxone, CXM Cefuroxime, CAZ Ceftazidime, ETP Ertapenem, MEM Meropenem, GM Gentamicin, AN Amikacin, CIP Ciprofloxacin, SXT Sulphamethoxazole – Trimethoprim, TET Tetracycline, NT Not tested

Overall, 89/150 (59.3%) of bacterial isolates from the ICU environment were MDR (Table 6). Among these isolates, P. aeruginosa 18/21 (85.7%) was the most dominant MDR followed by K.pneumoniae 16/22 (72.7%) and Acinetobacter spp., 8/11(72.7%). These bacterial isolates were resistant to at least three up to nine classes of antibiotic agents (Table 6).

Table 6.

Multidrug-resistant bacteria from intensive care unit environments at Addis Ababa, Ethiopia, 2024/2025

Gram –Negative Bacteria	Resistant antibiotic combination	No of drug classes	Frequency n (%)	MDR n (%)
E.coli (n = 15)	PIP, PTZ, FEP, CTX, CRO, CXM, CAZ, AT, ETP, MEM, GM, AN, CIP, SXT, TET	9	1(6.7)	11(73.3)
	PIP, PTZ, FEP, CRO, CAZ, AT, MEM, GM, AN, CIP, SXT	8	2(13.3)
	PIP, FEP, CTX, CRO, CXM, CAZ, AT, ETP, MEM, GM, AN, SXT, TET	7	1(6.7)
	PIP, FEP, CTX, CRO, CXM, CAZ, AT, MEM, GM, AN, CIP, TET	7	1(6.7)
	FEP, CRO, CAZ, AT, MEM, GM, AN, SXT, TET	6	3(20)
	PIP, FEP, CTX, CRO, CXM, CAZ, AT, MEM, GM, AN	5	2(13.3)
	PIP, FEP, TET	3	1(6.7)
K.pneumoniae (n = 22)	PTZ, FEP, CTX, CRO, CXM, CAZ, AT, ETP, MEM, GM, AN, CIP, SXT, TET	8	1(4.5)	16(72.7)
	PTZ, FEP, CTX, CRO, CXM, CAZ, AT, ETP, MEM, GM, CIP, SXT, TET	7	2(9.1)
	PTZ, FEP, CTX, CRO, CXM, CAZ, AT, ETP, MEM, GM, SXT	6	4(18.2)
	PTZ, FEP, CTX, CRO, CXM, CAZ, ETP, MEM, GM, TET	5	3(13.6)
	PTZ, CTX, CXM, SXT, TET	4	3(13.6)
	PTZ, CTX, CRO, CXM, CAZ, TET	3	3(13.6)
P.aeruginosa (n = 21)	PIP, PTZ, FEP, CTX, CRO, CAZ, AT, MEM, GM, AN, CIP, SXT, TET	9	1(4.7)	18(85.7)
	PIP, PTZ, FEP, CTX, CRO, CAZ, MEM, GM, CIP, SXT, TET	8	2(9.5)
	PIP, CTX, CRO, CAZ, MEM, GM, AN, CIP, SXT, TET	7	1(4.7)
	PTZ, FEP, CTX, CRO, CAZ, MEM, GM, SXT, TET	6	2(9.5)
	PTZ, FEP, CTX, CRO, CAZ, MEM, GM, SXT	5	3(14.3)
	PIP, CTX, CRO, CAZ, GM, CIP, TET	5	5(23.8)
	PIP, CTX, CRO, CAZ, GM, TET	4	3(14.3)
	PIP, PTZ, FEP, CAZ	3	1(4.7)
A.baumannii (n = 18)	PIP, PTZ, FEP, CTX, CRO, CAZ, AT, MEM, GM, AN, CIP, SXT, TET	9	2(11.1)	12(66.7%)
	PIP, PTZ, FEP, CTX, CRO, CAZ, MEM, GM, CIP, SXT, TET	8	4(22.2)
	PIP, FEP, CTX, CRO, CAZ, AT, MEM, GM, SXT, TET	7	1(5.6)
	PIP, PTZ, CTX, CRO, CAZ, AT, MEM, TET	6	1(5.6)
	PIP, PTZ, CTX, CRO, CAZ, TET	4	1(5.6)
	PIP, CTX, CRO, CAZ, SXT	3	1(5.6)
	PTZ, CAZ, GM, AN	3	1(5.6)
	PIP, CAZ, CIP	3	1(5.6)
Acinetobacter spp (n = 11)	PIP, PTZ, FEP, CTX, CRO, CAZ, AT, MEM, GM, AN, CIP, SXT, TET	9	2(18.2)	8(72.7)
	PIP, CAZ, GM, AN, CIP, SXT, TET	6	2(18.2)
	PIP, CTX, CRO, CAZ, AT, TET	4	1(9.1)
	PIP, CTX, CRO, CAZ, AT	3	1(9.1)
	PTZ, CAZ, TET	3	2(18.2)
E.cloacae (n = 9)	FEP, CTX, CRO, CAZ, MEM, CIP, SXT,	5	1(11.1)	3(33.3)
	FEP, CXM, AT, ETP, MEM	4	2(22.2)
K.oxytoca (n = 9)	PTZ, CAZ, CIP, TET	4	1(11.1)	4(44.4)
	FEP, CTX, CRO, CXM, CAZ, SXT, TET	3	1(11.1)
	CRO, SXT, TET	3	2(22.2)
Gram-Positive Bacteria
S. aureus (n = 14)	PEN, OXA/AMP, ERY, TET, CIP, NIT, CLO, SXT, GEN, FOX	9	1(7.1)	5(35.7)
	PEN, OXA/AMP, ERY, TET, CIP, NIT, SXT, GEN, FOX	9	1(7.1)
	ERY, TET, GEN, FOX	4	2(14.3)
	PEN, CLO, SXT	3	1(7.1)
Enterococcus spp (n = 12)	PEN, ERY, TET, VAN	4	2(16.7)	6(50)
	PEN, ERY, VAN	3	4(33.3)
S.epidermidis (n = 6)	PEN, ERY, TET, CIP, CLO, SXT, GEN	7	1(16.7)	3(50)
	PEN, TET, SXT	3	2(33.3)
S.haemolyticus (n = 8)	PEN, ERY, TET, CIP, SXT, GEN, FOX	7	1(12.5)	3(37.5)
	PEN, ERY, TET, CIP, SXT	5	1(12.5)
	PEN, CLO, SXT	3	1(12.5)
Total (n = 150)				89(59.3)

Abbreviations: PEN Penicillin, AMP Ampicillin, ERY Erythromycin, TET Tetracycline, CIP Ciprofloxacin, CLO Clindamycin, GEN Gentamycin, VAN Vancomycin, FOX Cefoxitin, PIP Piperacillin, PTZ Piperacillin/tazobactam, AMC Amoxicillin/clavulanic acid, FEP Cefepime, CTX Cefotaxime, CRO Ceftriaxone, CXM Cefuroxime, CAZ Ceftazidime, ETP Ertapenem, MEM Meropenem, AN Amikacin, SXT Sulphamethoxazole – Trimethoprim

Magnitudes of extended-spectrum beta-lactamase and carbapenemase- producing isolates

Among the 46 GNB enterobacterales isolates, E.coli (15), K. pneumoniae (22), and K.oxytoca (9), tested for ESBL production, 35/46 (76.1%) were presumptive ESBL producers. Among these, 35 potential ESBL-producing isolates, 23/35 (65.7%), were confirmed as ESBL producers. The overall magnitudes of ESBL producers among GNB isolates were 23/105 (21.9%). The distribution of ESBL-producing isolates among enterobacterales was 23/46 (50%), with the highest proportion in K. pneumoniae 10/46 (21.7%), and the lowest rate in K. oxytoca 4/46 (8.7%). Among ESBL producers, the highest rates were found in K. pneumoniae 10/23 (43.5%) and E. coli 9/23 (39.1%).

Of 105 GNB isolates, 36/105 (34.3%) were presumptive CP. Of these 36 potential CP isolates, 10/36 (27.8%) were confirmed to be CP. The overall distribution of CP among GNB isolates was 10/105 (9.5%), with the highest proportion in K. pneumoniae 4/105 (3.8%). There was no carbapenemase production in K.oxytoca. Among CP isolates, the most dominant species were K. pneumoniae 4/10 (40%), as exhibited in Fig. 1.

Fig. 1.

Magnitude of ESBLs and CP isolates in the Intensive Care Unit environments, Addis Ababa, Ethiopia, 2024/25

The highest proportion of ESBL-producing isolates from the ICU environment was found from the suction machine, 4/23 (17.4%), followed by linens, 3/23 (13%); whereas, the highest proportion of CP isolates was found from the patient bed, 4/17 (23.5%), linens, and suction machine, with a similar rate, 3/17 (17.6%). No CPs were detected on the walls near patient beds as depicted in Fig. 2.

Fig. 2.

Distribution of ESBL and CP GNB isolates over inanimate surfaces and medical devices in the Intensive Care Unit environments, Addis Ababa, Ethiopia, 2024/25

Discussion

This study presents a comprehensive analysis of the occurrence and AMR profiles of bacterial species recovered from ICU environments in two hospitals in Addis Ababa, Ethiopia. In contrast to previous studies that primarily focused on clinical isolates, our findings demonstrate a substantial environmental reservoir of AMR bacteria, including ESKAPE-E pathogens, ESBL-producing organisms, and CP bacteria within ICU settings. These findings highlight the critical importance of strengthening IPC measures in ICUs, alongside promoting the prudent use of antimicrobials, to reduce the dissemination and persistence of MDR bacteria.

In this study, 54.4% of 239 swab samples showed positive results for bacterial contamination. There were lower reports from Nigeria (39.4%) [19] and Sudan (29.7%) [20]. Nevertheless, a higher magnitude of bacterial isolates was obtained from Mekelle, Ethiopia (88.4%) [21] and Zimbabwe (86.2%) [22]. This discrepancy may be attributed to differences in infection prevention and control (IPC) practices, including ventilation systems, hand hygiene compliance, sterilization procedures, and disinfection protocols.

Both GPB (32.1%) and GNB (67.9%) were recovered from the ICU environments. Comparable findings were reported from Zimbabwe [22] and Addis Ababa, Ethiopia [3]. However, a higher magnitude of GPB was reported from Iran [23] and Gondar, Ethiopia [24]. The increased magnitudes of GNB isolates in our study may be attributed to several factors, including the presence of colonized patients, suitable environmental conditions that support the growth of GNB, lack of disinfection practices, various epidemiological influences, differences in sample size, and sampling techniques.

K. pneumoniae (13.6%), followed by P. aeruginosa (13%) was the most prevalent bacterial species. These findings aligned with reports from Gondar, Ethiopia, and Iran [23, 25]. However, inconsistent with reports from Addis Ababa, Ethiopia, where CoNS were the most frequently recovered bacterial isolates [26]. This discrepancy may be due to variation in sample size, study design, culture technique used, and geographical location.

Patient beds followed by suction machines were the most contaminated surfaces. There are similar studies comparable with this study in Ethiopia and global studies [25, 27]. P. aeruginosa (30%) was predominantly recovered from the suction machines and ventilators. Floor surfaces were mainly dominated by E. coli (20%), while sinks were predominantly colonized by B. cereus. (42.8%). These findings were also consistent with other studies from ICU environments [28, 29]. These discrepancies may be differences in IPC, disinfection practices, sample size, and geographical features. The high occurrence of P. aeruginosa on ventilators is associated with its biofilm-forming capacity via alginate production [30].

From the total isolates obtained in this study, ESKAPE-E pathogens comprised 68.5%. There is a very limited study regarding ESKAPE-E under this collective term from ICU environments in Ethiopia. However, there are reports from clinical isolates. A study from the USA reported 42.2% of ESKAPEE from bloodstream infection [31]. Whereas another study from Ethiopia reported 65.3% of bacterial species were ESKAPEE pathogens from surgical site infection [32]. Though it is difficult to compare environmental findings of ESKAPE-E with clinical studies, the presence of these group of pathogens in hospitalized patients may indicate their potential capacity to spread into the clinical environments.

In the present study, all S. haemolyticus were resistant to penicillin, and 8/12 (66.7%) of Enterococcus spp were resistant to erythromycin, tetracycline, and penicillin in GPB isolates. These findings were consistent with reports from Ethiopia [33, 34]. Whereas, all P. aeruginosa exhibited a high resistance rate to ceftazidime, and 86.5% and 81.8% of K.pneumoniae isolates were resistant to cefuroxime and cefotaxime, respectively, in GNB isolates. This finding is comparable with a report from Morocco [28]. Most bacterial isolates showed high susceptibility to ertapenem and amikacin. This result is supported by studies from Gondar, Ethiopia [34], and Debre Markos, Ethiopia [35]. The lower resistance rates of ertapenem may be associated with the intrinsic resistance nature of most of our isolates, like P.aeruginosa, A.baumannii, and Acinetobacter spp., and the lower resistance of amikacin may reflect its less frequent use in routine clinical practice in Ethiopia, as it was reserved for rifampicin-resistant or MDR-Tuberculosis (TB) [36].

Of the total bacterial isolates recovered, 59.3% exhibited MDR. P. aeruginosa exhibited the highest MDR rate (85.7%). Lower MDR rates were obtained from Debre Markos, Ethiopia [35], and Woliata Sodo, Ethiopia [37]. This discrepancy may be due to variation in sample size, study design, culture technique used, geographical location, and antimicrobial usage.

Among the total GNB isolates recovered, 21.9% were ESBL producers. Lower magnitudes of ESBL were reported in Debre Berhan, Ethiopia (14.5%) [38] and the United Kingdom (3.1%) [39]. However, a higher magnitude of ESBLs was reported from Thailand (54.8%) [40]. The observed discrepancy may be explained by variations in sampling locations, IPC measures, and geographical context. Furthermore, differences in patient occupancy and the prevalence of infected patients during the sampling period may have influenced the distribution and detection of bacterial isolates.

Among the total bacterial isolates recovered, 16.2% were CP isolates. These CP rates were notably higher than those reported in Debre Berhan, where 10.3% of isolates were obtained from healthcare worker gowns [38] and in Addis Ababa, Ethiopia, where 12.2% of bacterial isolates were CP from ICU and operating theater environments [26]. This discrepancy may be due to variation in sample size, study design, culture technique used, geographical location, and carbapenem antibiotic use. The occurrence of these CP bacterial isolates in both adult and neonatal ICU environments, particularly in the neonatal ICU, may be a crucial risk factor for neonatal infection in sub-Saharan countries where neonatal sepsis is the main cause of death of neonates [41]. Therefore, these findings emphasize the importance of adopting effective disinfection technologies to ensure a safer ICU environment and reduce the spread of resistant bacteria.

Previous studies from Ethiopia have reported the prevalence of CP bacterial isolates in clinical settings. In Addis Ababa, 2.6% of bacterial isolates were confirmed as CP producers [36], while a prevalence of 6% was reported in northwestern Ethiopia [42]. A systematic review further demonstrated considerable variation in the prevalence of CP isolates across different clinical contexts. For example, the prevalence among urinary tract infection (UTI) isolates in Addis Ababa was 2.73% [43], whereas a substantially higher prevalence of 30% was reported among surgical site infection isolates in southern Ethiopia [44]. The detection of CP isolates in clinical specimens suggests that these resistant pathogens may also be present within hospital environments, where contamination can occur through the shedding of bacteria by colonized or infected patients [45].

Limitations of the study

Sampling was limited to two hospitals in Addis Ababa, which may limit the generalizability of the results. It didn’t employ minimum inhibitory concentration (MIC) for the AST of bacterial isolates. Molecular methods and whole genome sequencing to detect resistance and virulence genes associated with ESBLs, CP, and their mechanisms were not conducted due to resource constraints. Finally, a longitudinal study was not conducted. Therefore, a genomics-based study with a longitudinal study design is needed in the future to understand the molecular diversity of bacteria in ICU settings.

Conclusion

This study underscores the high magnitude of resistant bacteria, particularly third-generation cephalosporin- and carbapenem-resistant A. baumannii, P. aeruginosa, and K. pneumoniae, in ICU settings. Many of the bacterial isolates also exhibited resistance to multiple antibiotic classes. The detection of these AMR isolates emphasizes the urgent need for stringent IPC measures.

Implementing advanced ICU environmen disinfection measures, routine AST to guide empirical treatment, and developing antimicrobial stewardship programs are crucial to curb the dissemination of superbugs. Addressing contamination on surfaces and medical equipment is vital to reducing the spread of MDR bacterial species, ultimately improving patient safety.

Authors' contributions

BMD: Conception, investigation, execution, software, writing original draft, analysis, and interpretation. BAK: Supervision, conception, writing, review, and editing. TAA: Supervision, conception, writing, review, and editing. AGB: Supervision, conception, writing, review, and editing. TST: Main supervision, conception, investigation, resource, writing, review, and editing.

Funding

Not applicable.

Data availability

All materials and data supporting these findings are included in this manuscript. The datasets used and /or analyzed during this study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol was reviewed and approved by the Addis Ababa University Institute of Biotechnology Research Ethics Review Committee (IoB-IRB) (Minute No. IoB/L-7/2016/2024) and the Addis Ababa Health Bureau Institutional Review Board (IRB). The study involved environmental sampling from intensive care unit (ICU) environments and did not include human participants and human specimens. Consequently, the principles of the Declaration of Helsinki were not applicable. The requirement for informed consent was also deemed not applicable as the study did not involve human subjects, in accordance with the National Research Ethics Review Guideline (NRERG, No. 01/2020, October 2020) of Ethiopia, which allows exemptions for minimal-risk research involving environmental sampling without direct human involvement. Permission to conduct the study was obtained from the administrative offices of Menelik II Comprehensive Specialized Hospital and Tirunesh Beijing General Hospital before sample collection.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.