Distribution of extended-spectrum β-lactamases (blaSHV, blaTEM, blaCTX-M) and AmpC β-lactamases -encoding genes among second- and third-generation cephalosporin-resistant Escherichia coli clinical isolates in Sana’a, Yemen

Abstract

Background

The global dissemination of extended-spectrum β-lactamases (ESBL) and AmpC β-lactamases among Escherichia coli poses a therapeutic challenge, contributing to the failure of empirical broad-spectrum antibiotic regimens. In conflict-affected and resource-limited settings like Yemen, the lack of molecular surveillance compromises infection control. This study aimed to determine the genotypic distribution of ESBL-, AmpC-, and dual-producing E. coli isolates resistant to second- and third-generation cephalosporins (specifically, ceftazidime, cefotaxime, or cefoxitin) across major health centers in Sana’a, Yemen.

Methods

A cross-sectional study was conducted enrolling 222 non-duplicate, consecutive clinical E. coli isolates that demonstrated phenotypic resistance or reduced susceptibility to second- and third-generation cephalosporins during routine diagnostic screening. Isolates were collected from Al-Yemen Al-Saeed Hospital, Modern European Hospital, and the National Center of Public Health Laboratories (NCPHL). Species confirmation was performed using standard biochemical profiling and the VITEK 2 automated system. Antimicrobial susceptibility testing followed CLSI criteria. Phenotypic production of ESBL and AmpC was validated using the Double-Disk Synergy Test (DDST) and boronic acid inhibition assays, respectively. Target resistance genes (blaSHV, blaTEM, blaCTX-M, blaMOX, blaCIT, blaDHA, blaACC, blaEBC, and blaFOX) were characterized via conventional multiplex polymerase chain reaction (PCR). Statistical analyses were performed using descriptive data and Pearson’s chi-square (χ²) tests.

Results

Of the 222 cephalosporin-resistant isolates, 112 (50.5%) were phenotypic ESBL-producers, 90 (40.5%) were AmpC-producers, and 20 (9.0%) were dual (ESBL+AmpC) co-producers. Anatomical specimen tropism was observed (χ² = 40.8, P = 0.001); ESBL-only producers significantly clustered in urinary tract infections (58.0%, adjusted residual = + 5.1), whereas AmpC-producers dominated wound infections (40.0%, adjusted residual = + 5.0). Genotypically, blaCTX-M was predominant in the ESBL group (107/112, 95.5%), while blaEBC (50.0%) and blaFOX (37.8%) were predominant in the AmpC group; blaEBC was also most common (55.0%) in dual-producing isolates. Carriage of at least one ESBL or AmpC gene was detected in 90.5% (201/222) of isolates. Phenotypic carbapenem resistance was observed in 14.4% of AmpC isolates (imipenem) and 20.0% of dual-producing isolates.

Conclusions

This study describes the distribution of ESBL and AmpC β-lactamases among clinical E. coli isolates resistant to second- and third-generation cephalosporins in Sana’a, Yemen, with blaCTX-M and blaEBC being the most frequently detected genes. A high proportion (90.5%) of isolates harbored at least one β-lactamase gene. The distinct associations of ESBLs with urine and AmpC with wounds support the need for targeted revisions to antibiotic stewardship. Furthermore, the detection of carbapenem resistance in AmpC and dual-producing strains highlights the need for molecular surveillance to monitor resistance trends in Yemeni clinical settings.

Introduction

Antimicrobial resistance (AMR) represents a significant global public health challenge, with β-lactamase-producing Escherichia coli being of particular concern. The World Health Organization has identified antimicrobial resistance as a threat requiring global action [1]. Extended-spectrum β-lactamase-producing Enterobacteriaceae have emerged as a public health concern worldwide, with E. coli being the most common species involved [2].

Extended-spectrum β-lactamases, primarily class A enzymes, confer resistance to penicillins, cephalosporins, and aztreonam but remain susceptible to carbapenems and β-lactamase inhibitors [3, 4]. The most clinically relevant ESBL families include TEM, SHV, and CTX-M types, with CTX-M enzymes becoming the most prevalent worldwide [5–7]. The global dissemination of CTX-M-type ESBLs has been widely documented, with these enzymes now detected in clinical and environmental isolates across all continents [6, 7].

AmpC β-lactamases (class C enzymes) represent another major resistance mechanism in E. coli [8]. These enzymes hydrolyze cephamycins in addition to penicillins and cephalosporins and are not inhibited by conventional β-lactamase inhibitors such as clavulanic acid [9]. The major plasmid-mediated AmpC families include MOX, CIT, DHA, ACC, EBC, and FOX, each with distinct epidemiological and biochemical characteristics [10]. The emergence of plasmid-mediated AmpC β-lactamases has complicated the treatment of infections caused by Gram-negative bacteria [9].

The co-occurrence of ESBL and AmpC genes in a single isolate further complicates resistance profiles and limits therapeutic options. This phenomenon has been increasingly reported in clinical settings, raising concerns about the potential for treatment failures and the need for more expensive and potentially toxic antimicrobial agents. The molecular characterization of these resistance determinants is important for understanding transmission dynamics and developing targeted intervention strategies [11].

Understanding the molecular epidemiology of β-lactamase genes is important for developing evidence-based treatment guidelines and infection control strategies [12, 13]. The prevalence and distribution of ESBL and AmpC genes vary across different geographical regions and healthcare settings [13]. However, comprehensive molecular data from many regions, including Yemen, remain limited. This study was undertaken to determine the prevalence and distribution of ESBL-encoding genes (blaSHV, blaTEM, blaCTX-M) and AmpC-encoding genes (blaMOX, blaCIT, blaDHA, blaACC, blaEBC, blaFOX) among clinical E. coli isolates and to analyze patterns of gene co-occurrence.

Materials and methods

Study design and sample collection

A cross-sectional study was conducted between October 2023 and December 2024 to determine the molecular epidemiology of β-lactamase-producing Escherichia coli. Clinical isolates were sequentially collected from the routine microbiology departments of two private hospitals (Al-Yemen Al-Saeed Hospital and Modern European Hospital) and the National Center of Public Health Laboratories (NCPHL) in Sana’a, Yemen.

A total of 222 non-duplicate positive cultures of E. coli isolates were enrolled.

Inclusion criteria

All positive cultures of E. coli isolates obtained from various specimens were enrolled; the inclusion criteria required isolates to exhibit phenotypic resistance or reduced susceptibility to second- and third-generation cephalosporins (specifically ceftazidime, cefotaxime, or cefoxitin) during routine diagnostic screening.

Sample size

The sample size was 192 E. coli positive cultures. The calculation was based on the E. coli prevalence of 14.6% (based on a study from Yemen [11]) and a 95% confidence interval margin of error of 5% (calculated using EPI Info version 7.1). To account for a 10% attrition rate, the final sample size reached 222 E. coli positive cultures.

Bacterial isolation and identification

All clinical specimens—including urine (n = 92), wound swabs (n = 51), pus (n = 38), sputum (n = 24), and blood (n = 17)—were processed according to standard clinical microbiological operating procedures [14]. Primary isolation was performed by inoculating specimens onto MacConkey agar and blood agar plates, followed by aerobic incubation at 37 °C for 18–24 h.

Presumptive phenotypic identification of E. coli was based on standard colony morphology (large, flat, pink, lactose-fermenting colonies on MacConkey agar), Gram-staining dynamics (Gram-negative rods), and standard biochemical profiling criteria aligned with MacFaddin manual guidelines [15]: indole test (positive), methyl red test (positive), Voges-Proskauer test (negative), citrate utilization test (negative), and urease production (negative). Species confirmation was performed using the VITEK 2 automated identification system (bioMérieux, Marcy-L’Etoile, France). Confirmed isolates were preserved in Tryptic Soy Broth (TSB) supplemented with 20% glycerol and stored at -80 °C for long-term preservation [14].

Antimicrobial susceptibility testing

Antimicrobial susceptibility testing was performed using the Kirby-Bauer disk diffusion method on Mueller-Hinton agar according to Clinical and Laboratory Standards Institute guidelines 2023 [16]. The following antibiotic disks were tested: amoxicillin-clavulanate (20/10 µg), ceftazidime (30 µg), ceftriaxone (30 µg), cefuroxime (30 µg), aztreonam (30 µg), ciprofloxacin (5 µg), gentamicin (10 µg), amikacin (30 µg), imipenem (10 µg), meropenem (10 µg), and colistin (10 µg). After incubation at 37 °C for 18–24 h, inhibition zone diameters were measured and interpreted as susceptible, intermediate, or resistant according to CLSI breakpoints. Escherichia coli ATCC 25,922 was used as a quality control strain. Isolates showing resistance or reduced susceptibility to second- or third-generation cephalosporins were considered potential ESBL or AmpC producers and were subjected to molecular testing [16].

Phenotypic screening and confirmation of ESBL and AmpC production

ESBL production

Initial screening was indicated by reduced susceptibility to ceftazidime or cefotaxime. Phenotypic confirmation was performed via the Double-Disk Synergy Test (DDST) as described by CLSI guidelines [16]. Ceftazidime (30 µg) and cefotaxime (30 µg) disks were placed at a distance of 25 mm (center-to-center) from an amoxicillin-clavulanic acid (20/10 µg) disk on a pre-inoculated Mueller-Hinton agar plate. A definitive enhancement or ghost zone expansion of the inhibition zone toward the amoxicillin-clavulanate disk confirmed ESBL production.

AmpC production

Screening was indicated by phenotypic resistance to cefoxitin (≤ 18 mm zone diameter). Confirmatory testing was conducted using the AmpC disk test with boronic acid as a specific AmpC inhibitor [17]. Isolates were tested against cefoxitin disks alone and cefoxitin disks supplemented with phenylboronic acid. An increase in the inhibition zone diameter of ≥ 5 mm for the cefoxitin-boronic acid combination compared to cefoxitin alone validated phenotypic AmpC production.

DNA extraction, quantification, and quality assessment

Bacterial genomic DNA was extracted using the boiling lysis method as previously outlined [18]. Briefly, 2–3 pure bacterial colonies were suspended in 200 µL of sterile nuclease-free water, boiled at 100 °C for 10 min in a thermal block, and immediately snap-cooled on ice for 5 min. The crude thermal lysate was centrifuged at 12,000 rpm for 5 min at 4 °C, and the clear supernatant containing the template DNA was transferred to sterile microcentrifuge tubes.

To assess template adequacy and extraction reproducibility across all 222 samples, DNA concentration and purity were measured using a NanoDrop spectrophotometer (Thermo Scientific, Waltham, MA, USA) at absorbance ratios of A₂₆₀/A₂₈₀. Samples exhibiting an A₂₆₀/A₂₈₀ ratio between 1.8 and 2.0 were considered suitable for downstream PCR kinetics. Genomic DNA was checked by running representative extracted DNA templates randomly on 0.8% agarose gel. Successful amplification of standard internal baseline target regions during subsequent analytical cycles served as a functional quality control check for template viability.

Conventional multiplex PCR optimization

Conventional polymerase chain reaction (PCR) assays were used to detect genes encoding core ESBL variants (blaSHV, blaTEM, blaCTX-M) and plasmid-mediated AmpC β-lactamase families (blaMOX, blaCIT, blaDHA, blaACC, blaEBC, blaFOX). Target genes, sequence-specific primers, anticipated amplicon sizes, and optimized primer annealing temperature (Tm) are shown in the Table below.

Primer sequences for detection of ESBL and AmpC β-lactamase genes

	Genes	Primer sequences (5′ − 3′)	Amplicon size (bp)	Tm	ACCESSION VERSION
ESBL	blaSHV	F- AGGATTGACTGCCTTTTTG R- ATTTGCTGATTTCGCTCG	393	53.42	MN477204
				53.89
	blaTEM	F- TTGGGTGCACGAGTGGGTTA R-TAATTGTTGCCGGGAAGCTA	504	61.70	AP045036
				57.21
	blaCTX-M	F- ACCGCCGATAATTCGCAGAT R-GATATCGTTGGTGGTGCCATAA	585	59.97	PX097631
				58.53
AmpC	blaFOX	F-AAC ATG GGG TAT CAG GGA GAT G R- CAA AGC GCG TAA CCG GAT TGG	190	59.29	MK733575
				62.76
	blaEBC	F-TCG GTA AAG CCG ATG TTG CGG R- CTT CCA CTG CGG CTG CCA GTT	302	63.59	CP188164
				65.80
	blaACC	F-AAC AGC CTC AGC AGC CGG TTA R- TTC GCC GCA ATC ATC CCT AGC	346	64.39	HG530658
				63.16
	blaDHA	F-AAC TTT CAC AGG TGT GCT GGG T R- CCG TAC GCA TAC TGG CTT TGC	405	62.61	LC542971
				62.24
	blaCIT	F-TGG CCA GAA CTG ACA GGC AAA R- TTT CTC CTG AAC GTG GCT GGC	462	62.48	AP023237
				63.18
	blaMOX	F- GCT GCT CAA GGA GCA CAG GAT R- CAC ATT GAC ATA GGT GTG GTG C	520	62.73	CP192748
				59.84
	16 S rRNA	F (5’-AGAGTTTGATCMTGGCTCAG-3’) R (5’-GGTTACCTTGTTACGACTT-3’)	1492	54.20 52.30	NR_119211

Amplifications were performed in a total reaction volume of 25 µL containing 12.5 µL of 2× PCR Master Mix (Qiagen, Germany), 1 µL of each forward and reverse primer (10 pmol/µL), 5 µL of extracted DNA template, and nuclease-free water to adjust volume profiles. Amplification protocols were performed in a thermal cycler under the following conditions: initial denaturation at 94 °C for 5 min; 35 cycles of denaturation at 94 °C for 30 s, primer annealing at gene-specific optimized temperatures (53.4–64.4 °C) for 30 s, and extension at 72 °C for 1 min; followed by a final terminal elongation at 72 °C for 10 min [17, 19].

Resulting amplicons were evaluated via horizontal electrophoresis using 1.5% (w/v) agarose gels stained with ethidium bromide in 1× TAE buffer. Bands were visualized under UV transillumination and sized using a 100 bp DNA ladder.

PCR quality control framework

No exogenous positive control was available for the ESBL and AmpC β-lactamase PCR assays. However, a subset of 20 randomly selected isolates was subjected to 16 S rRNA gene sequencing to confirm species identity. The successful amplification of the 16 S rRNA gene in all selected isolates served as an internal control to verify the quality of the extracted DNA and the efficacy of the PCR reagents and cycling conditions. Nuclease-free water was used as a negative control in every PCR run.

Statistical analysis

Data profiling was managed using SPSS software version 25.0 (IBM Corp., Armonk, NY, USA). Descriptive frequency trends were compiled to summarize baseline demographic sample sources, antimicrobial susceptibility layouts, and specific resistance gene distribution patterns. Inferential statistical associations involving categorical variables cross-tabulated across patient specimen groups were calculated using Pearson’s chi-square test. A P-value < 0.05 was considered statistically significant.

Results

Among the 222 E. coli isolates examined, 112 (50.5%) were ESBL-producers, 90 (40.5%) were AmpC-producers, and 20 (9.0%) co-produced both ESBL and AmpC enzymes. The distribution of ESBL, AmpC, and co-produced both ESBL and AmpC β-lactamase genes among these three groups is presented in Table 1.

Table 1.

distribution of extended-spectrum β-lactamases (blaSHV, blaTEM, blaCTX-M) and AmpC β-lactamases-encoding genes among Escherichia coli phenotypic groups

	Phenotypic Groups
Encoding Genes	ESBL (N=112)	AmpC (N=90)	ESBL and AmpC (N=20)
	No. (%)	No. (%)	No. (%)
blaSHV (n=10)	10 (8.9%)	0 (0%)	0 (0%)
blaTEM (n=74)	69 (61.6%)	0 (0%)	5 (25.0%)
blaCTX-M (n=117)	107 (95.5%)	0 (0%)	10 (50.0%)
blaMOX (n=7)	0 (0%)	7 (7.8%)	0 (0%)
blaCIT (n=12)	0 (0%)	11 (12.2%)	1 (5.0%)
blaDHA (n=6)	0 (0%)	6 (6.7%)	0 (0%)
blaACC (n=32)	0 (0%)	24 (26.7%)	8 (40.0%)
blaEBC (n=56)	0 (0%)	45 (50.0%)	11 (55.0%)
blaFOX (n=40)	0 (0%)	34 (37.8%)	6 (30.0%)

The present study revealed distinct patterns in the distribution of β-lactamase genes across the three phenotypic groups. In the ESBL group (N = 112), blaCTX-M was the most frequently detected gene (95.5%, 107/112).

Figure 1 illustrates the distribution between phenotypic resistance and the presence of specific encoding genes in Escherichia coli across the three groups. In the ESBL group (N = 112), 95.5% (107) of isolates tested positive for the corresponding genes, while 4.5% (5) did not. The AmpC group (N = 90) showed a detection rate of 85.6% (77 isolates), with 14.4% (13) testing negative. Among isolates exhibiting both phenotypes (N = 20), 85.0% (17) possessed the encoding genes, whereas 15.0% (3) did not.

Fig. 1.

Distribution of extended-spectrum β-lactamases (ESBL) and AmpC β-lactamases-encoding genes in Escherichia coli phenotypic groups

Overall, 201 of 222 isolates (90.5%) harbored at least one β-lactamase gene, while 21 isolates (9.5%) carried no detectable gene.

The number of β-lactamase genes per isolate varied across the three groups (Table 2). Among ESBL-producers, 5 isolates (4.5%) carried no detectable gene, 34 isolates (30.4%) carried one gene, 67 isolates (59.8%) carried two genes, 6 isolates (5.4%) carried three genes, and none carried four genes. Among AmpC-producers, 13 isolates (14.4%) carried no gene, 40 isolates (44.4%) carried one gene, 26 isolates (28.9%) carried two genes, 9 isolates (10.0%) carried three genes, and 2 isolates (2.2%) carried four genes. In the co-producing group, 3 isolates (15.0%) carried no genes, 4 isolates (20.0%) carried one gene, 5 isolates (25.0%) carried two genes, 5 isolates (25.0%) carried three genes, and 3 isolates (15.0%) carried four genes

Table 2.

Distribution of extended-spectrum β-lactamases and AmpC β-lactamases -encoding genes in Escherichia coli phenotypic groups by number among clinical isolates

Number of Genes	ESBL (N = 112)	AmpC (N = 90)	ESBL and AmpC (N = 20)
	No. (%)	No. (%)	No. (%)
No gene (21)	5(4.5%)	13(14.4%)	3(15.0%)
One (78)	34 (30.4%)	40 (44.4%)	4 (20.0%)
Two (98)	67 (59.8%)	26 (28.9%)	5 (25.0%)
Three (20)	6 (5.4%)	9 (10.0%)	5 (25.0%)
Four (5)	0 (0%)	2 (2.2%)	3 (15.0%)

The association between E. coli phenotypes and health units (Table 3) reveals a statistically significant relationship (χ² = 12.9, P = 0.002), indicating that the distribution of resistance phenotypes differs significantly between the laboratory and private hospital settings. The laboratory samples demonstrated a significantly higher prevalence of ESBL-producing E. coli (67.1%) compared to private hospitals (41.8%), supported by an adjusted residual of 3.6. In contrast, private hospital samples were significantly more likely to exhibit the AmpC phenotype (47.3%) compared to the laboratory (27.6%), with an adjusted residual of 2.8. The co-production of both ESBL and AmpC phenotypes did not show a statistically significant difference between the two groups, as the adjusted residuals (± 1.4) fell below the standard threshold for significance.

Table 3.

The association between E. coli phenotype and health unit

Groups		Health unit			Total	Χ2	P value
		Lab N = 76		Private hospital N = 146
ESBL	Count	51a	61b		112	12.9	0.002
	% within hospital	67.1%	41.8%		50.5%
	Adjusted Residual	3.6	-3.6
AmpC	Count	21a	69b		90
	% within hospital	27.6%	47.3%		40.5%
	Adjusted Residual	-2.8	2.8
ESBL and AmpC	Count	4a	16a		20
	% within hospital	5.3%	11.0%		9.0%
	Adjusted Residual	-1.4	1.4

Table 4 demonstrates a statistically significant association between the type of clinical specimen and the E. coli phenotypic resistance groups (χ² = 40.8, P = 0.001). The distribution of isolates varies notably by specimen source; urine samples are strongly associated with the ESBL group, accounting for 58.0% of those isolates compared to 21.1% in the AmpC group. Conversely, wound specimens show a strong association with the AmpC group, representing 40.0% of its isolates, whereas they make up only 9.8% of the ESBL group. Other specimens, including sputum, blood, and pus, exhibited no statistically significant associations.

Table 4.

The association between clinical specimens and Escherichia coli phenotypic groups

	Specimens	ESBL (N = 112)	AmpC (N = 90)	ESBL and AmpC (N = 20)	Total	X2	P value
sputum	Count	8a	14a	2a	24	40.8	0.001
	% within Groups	7.1%	15.6%	10.0%	10.8%
	Adjusted Residual	-1.8	1.9	− 0.1
urine	Count	65a	19b	8a, b	92
	% within Groups	58.0%	21.1%	40.0%	41.4%
	Adjusted Residual	5.1	-5.1	− 0.1
wound	Count	11a	36b	4a, b	51
	% within Groups	9.8%	40.0%	20.0%	23.0%
	Adjusted Residual	-4.7	5.0	− 0.3
Blood	Count	7a	8a	2a	17
	% within Groups	6.3%	8.9%	10.0%	7.7%
	Adjusted Residual	− 0.8	0.6	0.4
pus	Count	21a	13a	4a	38
	% within Groups	18.8%	14.4%	20.0%	17.1%
	Adjusted Residual	0.7	− 0.9	0.4

Table 5 provides a detailed breakdown of ESBL and AmpC gene detection across different specimen types. Among ESBL genes, the highest detection rates were observed in urine (63 positive, 68.5%), followed by pus (20 positive, 52.6%), blood (7 positive, 41.2%), sputum (6 positive, 25.0%), and wound (11 positive, 21.6%). For AmpC genes, wound specimens showed the highest detection rate (33 positive, 64.7%), followed by blood (8 positive, 47.1%), sputum (11 positive, 45.8%), pus (12 positive, 31.6%), and urine (13 positive, 14.1%).

Table 5.

Distribution of extended-spectrum β-lactamases and AmpC β-lactamases encoding genes in Escherichia coli phenotypic groups among clinical specimens

		Specimens (N = 222)
Phenotype groups	Gene	Wound (N = 51)	Urine (N = 92)	Sputum (N = 24)	Blood (N = 17)	Pus (N = 38)
ESBL (N = 112)	Positive n = 107	11(21.6%)	63(68.5%)	6 (25.0%)	7(41.2%)	20(52.6%)
	Negative N = 5	0 (0%)	2 (2.2%)	2 (8.3%)	0 (0%)	1 (2.6%)
AmpC (N = 90)	Positive N = 77	33(64.7%)	13 (14.1%)	11(45.8%)	8(47.1%)	12 (31.6%)
	Negative N = 13	3 (5.9%)	6 (6.5%)	3 (12.5%)	0 (0%)	1 (2.6%)
ESBL and AmpC (N = 20)	Positive N = 17	4 (7.8%)	6 (6.5%)	2 (8.3%)	2 (11.7%)	3 (7.9%)
	Negative N = 3	0 (0%)	2 (2.2%)	0 (0%)	0 (0%)	1 (2.6%)

The distribution of individual ESBL and AmpC genes across different clinical specimens is shown in Table 6. Among ESBL genes, blaCTX-M was most prevalent in urine (66 isolates, 71.7%), followed by pus (21 isolates, 55.3%), blood (9 isolates, 52.9%), sputum (8 isolates, 33.3%), and wound (13 isolates, 25.5%). blaTEM was also most common in urine (46 isolates, 50.0%), followed by pus (12 isolates, 31.6%), sputum (5 isolates, 20.8%), blood (3 isolates, 17.6%), and wound (8 isolates, 15.7%). blaSHV showed the lowest prevalence overall, with the highest detection in urine (6 isolates, 6.5%).

Table 6.

Distribution of extended-spectrum β-lactamases (blaSHV, blaTEM, blaCTX-M) and AmpC (blaMOX, blaCIT, blaDHA, blaACC, blaEBC, and blaFOX ) β-lactamases encoding genes in E. coli among clinical specimens

Gene	Total	Specimens (N = 222)
		Wound (N = 51)	Urine (N = 92)	Sputum (N = 24)	Blood (N = 17)	Pus (N = 38)
blaSHV	N = 10	1 (1.96%)	6 (6.5%)	1 (4.2%)	0 (0%)	2 (5.3%)
blaTEM	N = 74	8 (15.7%)	46 (50.0%)	5 (20.8%)	3 (17.6%)	12 (31.6%)
blaCTX-M	N = 117	13 (25.5%)	66 (71.7%)	8 (33.3%)	9 (52.9%)	21 (55.3%)
blaACC	N = 32	11 (21.6%)	6 (6.5%)	4 (16.7%)	4 (23.5%)	7 (18.4%)
blaEBC	N = 56	16 (31.4%)	13 (14.1%)	9 (37.5%)	7 (41.2%)	11 (28.9%)
blaFOX	N = 40	21(41.2%)	4 (4.3%)	6 (25.0%)	3 (17.6%)	6 (15.8%)
blaMOX	N = 7	4 (7.8%)	0 (0%)	0 (0%)	0 (0%)	3 (7.9%)
blaCIT	N = 12	4 (7.8%)	4 (4.3%)	4 (16.7%)	0 (0%)	0 (0%)
blaDHA	N = 6	3 (5.9%)	0 (0%)	2 (8.3%)	0 (0%)	1 (2.6%)

Among AmpC genes, blaFOX was most prevalent in wound (21 isolates, 41.2%), followed by sputum (6 isolates, 25.0%), pus (6 isolates, 15.8%), blood (3 isolates, 17.6%), and urine (4 isolates, 4.3%). blaEBC was most common in blood (7 isolates, 41.2%) and sputum (9 isolates, 37.5%). blaACC showed relatively high prevalence in blood (4 isolates, 23.5%) and wound (11 isolates, 21.6%). The less common AmpC genes (blaMOX, blaCIT, blaDHA) showed variable distribution, with generally lower prevalence rates.

The antimicrobial resistance profiles of ESBL-producing, AmpC-producing, and co-producing E. coli isolates are presented in Table 7. While some isolates showed phenotypic resistance to carbapenems, the investigation of carbapenemase-encoding genes was outside the scope of this study, which focused strictly on ESBL and AmpC molecular characterization.

Table 7.

The Antimicrobial resistance profiling of Escherichia coli phenotypic groups

Antibiotics	ESBL (N = 112)		AmpC (N = 90)		ESBL and AmpC (n = 20)
	Susceptible	Resistant	Susceptible	Resistant	Susceptible	Resistant
Colistin (CT)	112 (100.0%)	0 (0%)	90 (100.0%)	0 (0%)	20 (100.0%)	0 (0%)
Ciprofloxacin (CIP)	60 (53.6%)	52 (46.4%)	33 (36.7%)	57 (63.3%)	6 (30.0%)	14 (70.0%)
Meropenem (MR)	112 (100.0%)	0 (0%)	81 (90.0%)	9 (10.0%)	17 (85.0%)	3 (15.0%)
Cefepime (FEP)	112 (100.0%)	0 (0%)	23 (25.6%)	67 (74.4%)	5 (25.0%)	15 (75.0%)
Cotrimoxazole (COT)	52 (46.4%)	60 (53.6%)	22 (24.4%)	68 (75.6%)	6 (30.0%)	14 (70.0%)
Levofloxacin (LE)	70 (62.5%)	42 (37.5%)	41 (45.6%)	49 (54.4%)	7 (35.0%)	13 (65.0%)
Ceftazidime (CAZ)	0 (0%)	112 (100.0%)	13 (14.4%)	77 (85.6%)	0 (0%)	20 (100.0%)
Gentamicin (GEN)	72 (64.3%)	40 (35.7%)	42 (46.7%)	48 (53.3%)	5 (25.0%)	15 (75.0%)
Amoxicillin-clavulanic (AMC)	23 (20.5%)	89 (79.5%)	0 (0%)	90(100.0%)	0 (0%)	20 (100.0%)
Ceftriaxone (CTR)	0 (0%)	112 (100.0%)	14 (15.6%)	76 (84.4%)	0 (0%)	20 (100.0%)
Piperacillin-tazobactam (PIT)	56 (50.0%)	56 (50.0%)	34 (37.8%)	56 (62.2%)	3 (15.0%)	17 (85.0%)
Cefuroxime (CXM)	0 (0%)	112 (100.0%)	0 (0%)	90(100.0%)	0 (0%)	20 (100.0%)
Cefoperazone (CS)	0 (0%)	112(100.0%)	16 (17.8%)	74 (82.2%)	0 (0%)	20 (100.0%)
Amikacin (AK)	77 (68.8%)	35 (31.2%)	46 (51.1%)	44 (48.9%)	7 (35.0%)	13 (65.0%)
Tobramycin (TOB)	52 (46.4%)	60 (53.6%)	32 (35.6%)	58 (64.4%)	4 (20.0%)	16 (80.0%)
Imipenem (IPM)	112 (100.0%)	0 (0%)	77 (85.6%)	13 (14.4%)	16 (80.0%)	4 (20.0%)
Aztreonam (AZT)	104 (92.9%)	8 (7.1%)	46 (51.1%)	44 (48.9%)	10 (50.0%)	10 (50.0%)

Notably, ESBL-only producers (without AmpC co-production) demonstrated 100% susceptibility to meropenem and imipenem, and high susceptibility to cefepime. In contrast, carbapenem and cefepime resistance was predominantly observed among co-producing (ESBL+AmpC) isolates. High resistance rates were also observed for amoxicillin-clavulanic acid (79.5%), tobramycin (53.6%), cotrimoxazole (53.6%), and piperacillin-tazobactam (50.0%).

AmpC-producing isolates demonstrated 100% resistance to amoxicillin-clavulanic acid and cefuroxime, with high resistance rates to ceftazidime (85.6%), ceftriaxone (84.4%), cefoperazone (82.2%), cotrimoxazole (75.6%), and cefepime (74.4%). Notably, 14.4% of AmpC-producers showed resistance to imipenem, and 10.0% to meropenem, indicating emerging carbapenem resistance.

Co-producing isolates exhibited the highest resistance rates across most antibiotics tested, with 100% resistance to amoxicillin-clavulanic acid, ceftazidime, ceftriaxone, and cefuroxime. These isolates also showed high resistance to piperacillin-tazobactam (85.0%), tobramycin (80.0%), cefepime (75.0%), gentamicin (75.0%), ciprofloxacin (70.0%), cotrimoxazole (70.0%), levofloxacin (65.0%), and amikacin (65.0%). Carbapenem resistance was observed, with 20.0% resistance to imipenem and 15.0% to meropenem.

Table 8 presents the distribution of individual ESBL and AmpC genes according to the number of resistance genes carried by each isolate. Among ESBL genes, blaCTX-M was found in isolates carrying one to four genes, with the highest proportion (59.0%) in isolates carrying two genes. blaTEM was predominantly found in isolates with two genes (87.8%), while blaSHV was most common in isolates carrying three genes (60.0%).

Table 8.

Distribution of E. coli Extended-Spectrum β-lactamases (ESBL) and AmpC-Encoding Genes among Cases by Number

Genes
	One (N = 78)	Two (N = 98)	Three (N = 20)	Four (N = 5)
blaSHV (n = 10)	0 (0%)	4 (40.0%)	6 (60.0%)	0 (0%)
blaTEM (n = 74)	0 (0%)	65 (87.8%)	6 (8.1%)	3 (4.1%)
blaCTX-M (n = 117)	34 (29.1%)	69 (59.0%)	11 (9.4%)	3 (2.6%)
blaACC (n = 32)	3 (9.4%)	14 (43.8%)	12 (37.5%)	3 (9.4%)
blaEBC (n = 56)	21 (37.5%)	18 (32.1%)	12 (21.4%)	5 (8.9%)
blaFOX (n = 40)	18 (45.0%)	13 (32.5%)	5 (12.5%)	4 (10.0%)
blaMOX (n = 7)	0 (0%)	5 (71.4%)	1 (14.3%)	1 (14.3%)
blaCIT (n = 12)	2 (16.7%)	5 (41.7%)	5 (41.7%)	0 (0%)
blaDHA (n = 6)	0 (0%)	3 (50.0%)	2 (33.3%)	1 (16.7%)

Among AmpC genes, blaEBC showed the most diverse distribution, being present in isolates carrying one to four genes, with the highest proportion (37.5%) in single-gene isolates. blaFOX was most prevalent in single-gene isolates (45.0%), while blaACC was most common in isolates carrying two genes (43.8%). The less prevalent AmpC genes (blaMOX, blaCIT, blaDHA) showed variable distribution patterns, with blaMOX predominantly found in isolates with two genes (71.4%).

Discussion

This study provides a molecular characterization of ESBL and AmpC β-lactamase genes in clinical E. coli isolates, revealing the distribution patterns of these genes. The clinical isolates were gathered from various healthcare environments, including private hospitals and a centralized national reference laboratory. This selection of collection sites allowed for the representation of resistance profiles across different settings.

Our findings are consistent with several recent studies reporting gene distribution patterns across different resistance types. A study in Iraq reported that among uropathogenic E. coli isolates, CTX-M was the most common ESBL gene (76%), followed by TEM (75%) and SHV (73%) [20]. These findings are broadly similar to our distribution, where blaCTX-M was most prevalent (95.5%), followed by blaTEM (61.6%) and blaSHV (8.9%). For AmpC-producing isolates, a study in Mozambique reported that among AmpC-positive E. coli (33% of 75 isolates), blaFOX (n = 24) and blaMOX (n = 22) were most prevalent, followed by blaDHA (n = 13) and blaCMY (n = 1) [21]. This aligns in part with our AmpC group, where blaFOX (37.8%) and blaMOX (7.8%) were detected, although blaEBC (50.0%) was the most common in our study. Additionally, a study in the Netherlands demonstrated that ESBL and AmpC co-producers exhibited diverse resistance profiles, including ESBL genes (blaCTX-M 50.0%, blaTEM 25.0%) and AmpC genes (blaEBC 55.0%, blaACC 40.0%, blaFOX 30.0%) [22], which is comparable to the gene distribution observed in our co-producing group.

Conversely, some studies have reported different gene distribution patterns. A study in Nigeria found that blaTEM was the most prevalent resistance determinant (52.5%) among ESBL-producing E. coli isolated from diarrheic stools [23], whereas in our ESBL group, blaCTX-M was predominant (95.5%) and blaTEM was secondary (61.6%). For AmpC genes, a study in South Korea reported that CTX-M-55 was dominant (69.6%) while CMY-2 represented only 4.3% of AmpC producers among ESBL/AmpC-producing E. coli from swine production [24]. This differs from our AmpC group, where blaEBC (50.0%), blaFOX (37.8%), and blaACC (26.7%) predominated. These differences suggest variability in β-lactamase gene distribution across different populations and settings.

An observation in this study was the discordance between phenotypic and genotypic detection. Among the 112 phenotypically confirmed ESBL-producers, 5 isolates (4.5%) had no detectable ESBL gene by PCR. A possible explanation is that phenotypic resistance in these isolates could be mediated by alternative ESBL gene families not included in our multiplex PCR screening panel, such as OXA-1, PER, or VEB. Similarly, among 90 phenotypically confirmed AmpC-producers, 13 isolates (14.4%) had no detectable plasmid-mediated AmpC gene. This may be explained by the fact that E. coli possesses a chromosomal ampC gene; phenotypic hyperproduction in these isolates may result from mutations in the chromosomal promoter/attenuator regions causing endogenous gene overexpression rather than acquisition of plasmid-mediated AmpC genes [25].

Our results are consistent with studies reporting multi-gene carriage patterns. A study in Mozambique reported that among AmpC-producing E. coli isolates, 33% tested positive for one or more plasmid-mediated AmpC genes, with isolates containing two (n = 15) or three (n = 10) AmpC genes [21], supporting our observation that AmpC isolates can carry multiple genes.

However, other studies have reported different patterns. A study in Iraq reported that 67% of ESBL-producing E. coli isolates harbored two genes (blaTEM and blaSHV), 65% carried two genes (blaTEM and blaCTX-M-1), and 42% carried three genes (blaTEM, blaSHV, and blaCTX-M-1) [26]. A study in Iran reported that the most prevalent resistance gene profiles were blaTEM alone (44.3%) followed by blaTEM/blaCTX-M (37.7%), indicating that single-gene and two-gene carriage patterns dominated their ESBL group [27]. These differences may be attributed to regional variations in antimicrobial resistance gene prevalence, differences in local antibiotic prescribing practices, and distinct plasmid transfer dynamics.

Our results show a statistically significant association between resistance phenotypes and healthcare facility type (P = 0.002). ESBL isolates were more prevalent at the central laboratory (NCPHL) (67.1%) compared to private hospitals (41.8%), while AmpC isolates were more prevalent in private hospitals (47.3%) compared to the central laboratory (27.6%). No significant difference was observed for dual producers (ESBL+AmpC).

The higher prevalence of ESBL at the central laboratory may reflect community-acquired infections referred from primary care settings, where broad-spectrum cephalosporin use is common [2]. In contrast, the higher AmpC prevalence in private hospitals may be related to more frequent surgical procedures, catheter use, and healthcare-associated infections that exert selective pressure from cephamycin and amoxicillin-clavulanate [3]. These findings are consistent with studies showing that resistance pattern distribution varies across different healthcare settings based on local antibiotic prescribing practices and patient populations [28]. The similar distribution of dual producers across both facility types suggests these isolates may be disseminated via mobile genetic elements that spread independently of facility type [29].

These findings suggest that resistance patterns vary by healthcare facility, which may support the development of customized antibiotic stewardship and infection control policies tailored to each setting [1].

A study from India reported that 88% (22/25) of AmpC-producing E. coli isolates were ESBL and AmpC co-producers, with predominance of blaTEM (n = 20), and co-carriage of blaCMY-2 and blaTEM (n = 15) in transconjugants [30]. This is comparable to our co-producing group, which showed multi-gene carriage (25.0% with three genes and 15.0% with four genes). However, other studies have reported different patterns. A study from South Korea reported that all ESBL/AmpC-producing E. coli isolates carried only one type of CTX-M β-lactamase or CMY β-lactamase, with no isolates carrying more than one genotype [24]. This differs from our findings, where two-gene carriage predominated in the ESBL group (59.8%), and 28.9% of AmpC isolates carried two genes, and 10.0% carried three genes.

A study from India found that among urinary E. coli isolates, 31% were ESBL producers, 62% were AmpC producers, and 7.04% were ESBL and AmpC co-producers [31]. This is consistent with our finding that urine specimens contained substantial proportions of all three resistance types.

In contrast, a study from Indonesia reported that the highest prevalence of ESBL-producing E. coli was found in pus (35.6%) and urine (27.2%) [32]. This differs from our ESBL group, where urine specimens were the predominant source, followed by wound specimens.

A study from Bangladesh reported that among 150 E. coli isolates, urine specimens (n = 132) yielded the highest number of ESBL-positive (28.0%), AmpC-positive (16.7%), and co-producing isolates (4.5%), followed by pus specimens [33]. This is consistent with our observation that urine and pus specimens were common sources of β-lactamase-producing E. coli.

A study from Iraq reported that blaCTX-M (70%) and blaTEM (52%) were prevalent ESBL genes in urine specimens [34]. Similarly, a study from Mozambique found that blaCTX-M (77%) was the most common ESBL gene, while blaFOX was frequently detected in urine and blood specimens [21]. These findings are consistent with our observation of high CTX-M prevalence in urine and blood specimens.

In contrast, a study from Ghana reported blaFOX (64%) and blaEBC (27%) as the predominant AmpC genes in clinical specimens, including urine, wounds, and blood [35]. This differs somewhat from our finding, where blaEBC and blaFOX were also common but with different proportions, suggesting geographic variation in AmpC gene distribution.

Our findings are consistent with studies reporting distinct antimicrobial resistance patterns across different β-lactamase-producing E. coli groups. A study from the Netherlands compared resistance profiles for AmpC, ESBL, and co-producing E. coli strains [22], supporting our observation of distinct patterns among these three phenotype groups. Similarly, a study from Egypt found that ESBL-producing Enterobacteriaceae showed 84.85% resistance to cefotaxime and ceftazidime, 60.61% resistance to levofloxacin, and 9.1% resistance to imipenem [36], comparable to our finding that ESBL producers exhibited high β-lactam and fluoroquinolone resistance with relatively lower carbapenem resistance.

A study from Brazil reported that ESBL-producing E. coli showed 85% resistance to cefotaxime, 74% to aztreonam, 81% to cefepime, 66% to amoxicillin-clavulanate, and 66% to ciprofloxacin, while AmpC-producing isolates demonstrated 100% resistance to amoxicillin-clavulanate and cefoxitin but lower fluoroquinolone resistance (39% to ciprofloxacin) [37]. These findings reflect resistance profiles similar to those observed between our ESBL and AmpC groups.

A study from Pakistan found that among ESBL-producing genes, blaCTX-M was the most prevalent (91.2%), followed by blaTEM (61.8%) and blaSHV (17.6%), with the most common gene combination being blaCTX-M + blaTEM (30.5%) [38]. This is consistent with our finding that CTX-M and TEM were common in ESBL-producing E. coli.

Limitations

A limitation of this study is the absence of commercially validated reference strains for PCR-positive controls for the target resistance genes. Consequently, amplicon identity for the β-lactamase PCR relied on size comparison with a DNA ladder. While this is an accepted approach, a subset of isolates was independently validated via 16 S rRNA gene sequencing, confirming both DNA quality and species identification. However, the possibility of non-specific amplification in the resistance-gene PCR cannot be fully excluded.

Conclusions

This study provides molecular insights into the distribution of ESBL and AmpC β-lactamases among clinical E. coli isolates in Sana’a, Yemen. The findings revealed a high prevalence of β-lactamase gene carriage, with 90.5% of isolates harboring at least one β-lactamase gene. Genotypically, blaCTX-M was the most frequently detected gene among ESBL-producing isolates, while blaEBC and blaFOX were the predominant plasmid-mediated AmpC families. The institutional distribution of these resistance phenotypes was similar across healthcare centers, suggesting widespread circulation within the local healthcare system. Their distribution across clinical specimens varied according to specimen type. Additionally, carbapenem resistance was detected in phenotypic AmpC and dual (ESBL+AmpC) co-producing isolates, a finding that warrants further monitoring and may pose challenges to treatment options in resource-limited settings.

Abbreviations

AMR

Antimicrobial Resistance

CLSI

Clinical and Laboratory Standards Institute

DDST

Double-Disk Synergy Test

DNA

Deoxyribonucleic Acid

ESBL

Extended-Spectrum β-Lactamase

MDR

Multidrug Resistance

NCPHL

National Center of Public Health Laboratories

PCR

Polymerase Chain Reaction

UTI

Urinary Tract Infection

Authors’ contributions

A.M.M.A. wrote the main manuscript. A.Y.A. supervised and S.S.B. reviewed. A.M.M.A., A.A., W.Y.A., performed the statistical analysis, data collection, and laboratory work. and R.A., F.A.B.A., and S.M.A. laboratory work.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of the Faculty of Medicine and Health Sciences, Sana’a University (approval number (26) 13 − 2) and by the National Center of Public Health Laboratories (NCPHL) and the participating hospitals in Sana’a, Yemen. All procedures performed in this study involving humans/human data/the use of human samples were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The requirement for informed consent was waived by the Ethics Committee of the Faculty of Medicine and Health Sciences, Sana’a University because the study used only de-identified bacterial isolates collected as part of routine clinical care, with no access to patient identifiers.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.