Dissemination of Extended-Spectrum β-Lactamase-Producing Enterobacterales in Organic Fertilizers: A One Health Perspective from Southwestern Colombia

Abstract

Extended-spectrum β-lactamase (ESBL)-producing bacteria are a growing public health concern within the One Health framework. This study aimed to characterize ESBL-producing Enterobacterales in industrial and artisanal organic fertilizers marketed in southwestern Colombia. Five commercial fertilizer brands were analyzed using a selective culture on ceftriaxone supplemented media (4 µg/mL), antimicrobial susceptibility testing by broth microdilution to determine minimum inhibitory concentrations (MICs), phenotypic synergy testing for ESBL confirmation, and polymerase chain reaction (PCR) to detect bla_TEM, bla_SHV, and bla_CTX-M genes. Overall, 18.6% of the samples showed growth of ceftriaxone-resistant Enterobacterales, predominantly Escherichia coli and Klebsiella pneumoniae. ESBL producers accounted for 84% of the isolates, all of which carried at least one bla gene, predominantly bla_CTX-M. Statistically significant differences in bacterial growth frequency were observed among fertilizer types, with higher positivity rates observed in manure-based artisanal formulations (p < 0.05). Whole-genome sequencing of selected isolates identified Klebsiella pneumoniae ST37 and Escherichia coli ST224, both harboring bla_CTX-M-55 and additional resistance and virulence determinants. These findings demonstrate that organic fertilizers, particularly manure-derived products, may act as reservoirs and potential dissemination routes for clinically relevant antimicrobial-resistant bacteria. This is the first study in Colombia documenting the presence of ESBL-producing bacteria in organic fertilizers. These results underscore the need to incorporate surveillance of these products into national policies under a One Health perspective.

1. Introduction

While fertilizers are vital for soil fertility and agricultural productivity, most conventional types are derived from mining or industrial compounds, such as petroleum derivatives or ammonium nitrate, which can release toxic substances into the soil [1]. These environmental drawbacks have encouraged the development of more sustainable alternatives, such as organic fertilizers composed of plant, animal, and mineral-based products. Because of their natural origin, these materials are more easily metabolized by plants, thereby improving soil quality, reducing greenhouse gas emissions, and increasing the organic matter content in soil [2].

In Colombia, the production of organic fertilizers is regulated by the national technical standard NTC 5167, which allows the inclusion of animal- or plant-based raw materials, as well as biosolids and residues from domestic or agro-industrial wastewater [3]. However, this regulation overlooks the widespread use of antimicrobial agents in livestock and agricultural practices. Antibiotics are commonly employed as growth promoters in animal production systems, which has led to the emergence of resistant bacteria within the intestinal microbiota. The incomplete metabolism and excretion of these compounds into the environment promote the persistence of active antimicrobial residues, which facilitate the selection and spread of resistance genes across environmental microorganisms [4].

Once these resistant bacteria enter the environment, they can cause infections in humans through direct contact with manure or through exposure to contaminated water and aerosols generated during fertilizer application, eventually entering the food chain [4]. Acknowledging such global concerns, the World Health Organization (WHO) emphasized the need for a comprehensive approach to study antimicrobial resistance (AMR) under the One Health framework, which recognizes the interconnection between human, animal, and environmental health [5,6]. Research on AMR has focused mainly on β-lactamase-producing bacteria, particularly those carrying extended-spectrum β-lactamases (ESBLs), which hydrolyze penicillins, third-generation cephalosporins, and aztreonam [7]. They predominantly belong to order Enterobacterales and are associated with treatment failures and increased risk of clinical complications [8]. The genes encoding β-lactamases are mostly plasmid-borne, facilitating their horizontal dissemination. The most globally prevalent ESBL families produce CTX-M-, TEM-, and SHV-type enzymes, with CTX-M producers being the most widely distributed [9]. The presence of ESBL-producing Enterobacterales in the environment has been reported globally, particularly in regions with intensive agricultural and livestock antibiotic use [10]. ESBL-producing isolates have been detected in pig farm manure and sludge in Central Europe [11], in vegetables irrigated with untreated wastewater in Nigeria [12] and the United States [13]. Similarly, ESBL-producing isolates have been reported in vegetables and soils augmented with manure or organic fertilizers in Tunisia [14] and France [15].

In Colombia, β-lactam resistance has primarily been investigated in the clinical setting, with the first reports of ESBL-producing K. pneumoniae and E. coli isolated from nosocomial infection cases [16,17]. From a One Health perspective, Aristizabal et al., identified cephalosporin-resistant Enterobacterales strains in wastewater treatment plants, while Mondragon et al. identified landfill leachates as potential reservoirs of resistance genes [18].

One of the key relevant aspects of this research lies in the limited information currently available regarding ESBL-producing bacteria in agricultural inputs in Colombia. To date, no studies in Colombia have examined the presence of ESBL-producing bacteria in organic fertilizers. This topic is closely aligned with the One Health framework, as organic fertilizers may act as an interface connecting animal waste, environmental dissemination, agricultural soils, food production systems, and potential human exposure. Given the interrelationships together with the increasing use of organic fertilizers, this study analyzed ESBL-producing Enterobacterales isolates recovered from organic fertilizers marketed in southwestern Colombia and characterized their genetic context and enzyme types.

2. Materials and Methods

2.1. Type of Study and Sample Size

Five types of organic fertilizers were selected based on market availability, including three industrially processed products and two artisanal formulations. A non-probabilistic judgmental sampling design was used to define the sample size (n = 75), ensuring technical feasibility and analytical depth. A significance level of 0.05 was considered [19]. Within each fertilizer bag, sampling was performed using a grid-based random strategy to reduce spatial heterogeneity bias. First, a virtual grid was superimposed onto the fertilizer surface. Then, ten grid squares were randomly selected using Microsoft Excel. Approximately 1 g of material was collected from each square for a total of 10 g. For each fertilizer bag, three independent composite samples were prepared using the same procedure (Figure 1).

Sample identification followed an alphanumeric system. For instance, the code FA1A, indicates that the sample is a fertilizer (F), and the second letter (A–C for industrial fertilizers and D–F for artisanal fertilizers) denotes the fertilizer type. The number represents the bag number (1–5), and the final letter indicates the replicate (A, B, or C) (Figure 1). Of the two artisanal fertilizers, one (FE) contains humus, fruit peels, and eggshells, while the second (FD) is a mixture of black soil, cow manure, rice husks, and decomposing twigs and leaves.

2.2. Total Colony Count, Screening of Ceftriaxone-Resistant Enterobacterales (AXO-Resistant Enterobacterales) and Statistical Analysis

For the microbiological analysis, 10 g of each composite sample was homogenized in 90 mL of nutrient broth with vigorous manual agitation to obtain a 1:10 suspension, followed by incubation at 37 °C for 24 h as a non-selective enrichment step. After incubation, the suspensions were thoroughly mixed and serial tenfold dilutions were prepared. In order to estimate total recoverable Enterobacterales, aliquots (150 µL) from dilutions expected to yield countable colonies (30–300 CFU) were plated onto MacConkey agar (bioMérieux, Marcy l’Étoile, France) and incubated at 37 °C for 24 h.

The same dilution series obtained after non-selective enrichment was used for a selective enrichment of presumptive AXO-resistant Enterobacterales. Briefly, 10 µL from each dilution were inoculated into 990 µL of MacConkey broth (Merck, Darmstadt, Germany) supplemented with 4 µg/mL ceftriaxone (AXO) (Santa Cruz Biotechnology, Dallas, TX, USA) and incubated at 37 °C for 24 h. By following selective enrichment process, aliquots were plated onto MacConkey (bioMérieux, Marcy l’Étoile, France) agar under the same incubation conditions. Since enumeration was performed after enrichment steps, the CFU values were interpreted as semi-quantitative estimates of recoverable Enterobacterales rather than absolute bacterial counts.

In order to obtain a statistical description, the data collected were summarized using mean values and standard deviations (Supplementary Table S1). As CFU counts did not meet assumptions of normality, differences among OF fertilizer types were assessed using the Kruskal–Wallis test. When statistically significant differences were observed (p < 0.05), Dunn’s post hoc test was applied for pairwise comparisons. Statistical analyses were performed using R software version 4.5.3 [20].

2.3. Bacterial Identification

Representative colonies displaying distinct lactose-fermenting morphologies consistent with Enterobacterales were selected and subcultured for further characterization (identification and susceptibility test). To implement the procedure, pure isolates were subcultured onto blood agar plates and incubated at 37 °C for 24 h prior to identification. Bacterial identification was performed using GN identification cards in the VITEK^® 2 Compact system (bioMérieux, Marcy l’Étoile, France), following the manufacturer’s instructions. Bacterial suspensions were prepared and adjusted to 0.5 McFarland using a DensiCHEK™ instrument, after which they were loaded into the VITEK^® system for automated identification (bioMérieux, Marcy l’Étoile, France). Identification results were interpreted according to the manufacturer’s recommendations.

2.4. Determination of Minimum Inhibitory Concentration by Broth Microdilution and Phenotypic Detection of ESBLs

Minimum inhibitory concentrations (MICs) were determined using the broth microdilution method with Sensititre panels (TREK Diagnostic Systems, Westlake, OH, USA), according to the manufacturer’s instructions. The antibiotics evaluated were cefotaxime (FOT), ceftriaxone (AXO), ceftazidime (TAZ), cefepime (FEP), cefoxitin (FOX), piperacillin/tazobactam (P/T4), doripenem (DORI), imipenem (IMI), meropenem (MERO), ertapenem (ETP), ceftolozane/tazobactam (C/T), and ceftazidime/avibactam (CZA). Results were interpreted according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [21]. E. coli ATCC 25922 was used as the quality control strain.

Phenotypic detection of ESBL production was performed using the double-disk synergy test as described by Calvo et al. [22]. Briefly, a bacterial suspension adjusted to 0.5 McFarland in 0.85% saline solution was inoculated onto Mueller–Hinton agar plates. An amoxicillin/clavulanic acid (AMC) (Becton, Dickinson and Company, Sparks, MD, USA) disk was placed at the center of the plate, while cefotaxime (30 µg) and ceftazidime (30 µg) disks (Becton, Dickinson and Company, Sparks, MD, USA) were positioned 20 mm apart on either side of the AMC disk. Plates were incubated at 35 ± 2 °C for 24 h. ESBL production was indicated by an enhanced inhibition zone (“keyhole effect”) between the cephalosporin disks and the AMC disk, indicating synergy. The absence of this effect was interpreted as a negative result.

2.5. Detection of bla_TEM, bla_SHV, and bla_CTX-M.group1 Genes

All isolates were analyzed by polymerase chain reaction (PCR) to detect resistance genes. Each target gene was amplified in separate PCR reactions using specific primer sets. The primers and conditions used for detection were performed under the protocols previously described [23,24,25]. Amplifications were conducted on a SelectCycler™ II Thermal Cycler (Select BioProducts, Edison, NJ, UEA). PCR conditions were optimized individually for each primer pair, as shown in

Table 1. Primers sequences and PCR conditions used in this study.

Gene	Primer	Oligonucleotide Sequence (5′→3′)	PCR Conditions	Reference
blaSHV	SHVC F	5′-CGCCGGGTTATTCTTATTTGTCGC-3′	1 cycle of 5 min at 94 °C; 30 cycles of 1 min at 95 °C, 1 min at 51 °C, 1 min at 72 °C; final extension of 10 min at 72 °C	[23]
	SHVC R	5′-TCTTTCCGATGCCGCCGCCAGTCA-3′
blaTEM	TEM F	5′-CTTCCTGTTTTTGCTCACCCA-3′	1 cycle of 5 min at 94 °C; 30 cycles of 1 min at 94 °C, 1 min at 52 °C, 1 min at 72 °C; final extension of 10 min at 72 °C	[24]
	TEM R	5′-TACGATACGGGAGGGCTTAC-3′
blaCTX-M	CTX-M G1F	5′-CCCATGGTTAAAAAACACTGC-3′	1 cycle of 5 min at 95 °C; 30 cycles of 1 min at 95 °C, 1 min at 60 °C, 1 min at 72 °C; final extension of 10 min at 72 °C	[25]
	CTX-M G1R	5′-CAGCGCTTTTGCCGTCTAAG-3′

.

2.6. Whole-Genome Sequencing

Three isolates were selected for whole-genome sequencing: two K. pneumoniae isolates (one recovered from FD fertilizer containing bovine manure and one from industrially produced FA fertilizer) and one E. coli isolate obtained from FD fertilizer. Because the antimicrobial resistance profiles of the isolates were similar, representative strains were selected based on the type of fertilizer from which they were recovered.

Genomic DNA was extracted from overnight cultures using the DNeasy UltraClean Microbial Kit (Qiagen, Hilden, Germany a) following the manufacturer’s instructions. Whole-genome sequencing (WGS) was performed on an Illumina NextSeq 1000 platform (Illumina Inc., San Diego, CA, USA). Raw sequencing reads were assembled de novo using SPAdes with default parameters. Gene prediction and annotation were performed using the Bacterial and Viral Bioinformatics Resource Center (BV-BRC) platform.

Multi-locus sequence typing (MLST), as well as resistome and virulome analyses, were conducted using the Solugenomics platform based on the PubMLST database (v2.23.0) [26]. Capsular typing (wzi gene), and virulence gene typing using Kleborate for K. pneumoniae [27]. The Whole Genome project has been deposited in DDBJ/ENA/GenBank under the accession number PRJNA1348518.

3. Results

3.1. Colony Counts and Screening of AXO-Resistant Enterobacterales

Enterobacterales growth was detected in 18 out of 75 analyzed samples (24%) following enrichment procedures. Total recoverable Enterobacterales counts varied markedly among fertilizer types as shown in

Table 2. Comparison of total and ceftriaxone-resistant Enterobacterales counts across five fertilizers. Statistical significance is indicated by * (p < 0.05). AXO, ceftriaxone.

Types of Fertilizers	Total Enterobacterales			AXO-Resistant Enterobacterales
	Median	[Q1–Q3]	Kruskal–Wallis (p-Value)	Median	[Q1–Q3]	Kruskal–Wallis (p-Value)
A	7.5 × 1012	0–2.12 × 1013	* 0.001	2.0 × 107	0–1.0 × 108	* 0.001
B	0	0-0		0	0-0
C	0	0-0		0	0-0
D	7.0 × 1011	5.6 × 109–1.4 × 1012		8.0 × 107	3.0 × 107–1.0 × 108
E	0	0-0		0	0-0

. The highest counts were consistently observed in fertilizer D, followed by fertilizers A and E. In contrast, no detectable growth was observed in fertilizers B and C under the experimental conditions used. A similar pattern was observed for AXO-resistant Enterobacterales.

Statistically significant differences in microbial load were observed among the evaluated fertilizers (p = 0.001). For the total Enterobacterales count fertilizers A and D showed the highest median values, in contrast no growth was detected in groups B, C, and E. Similarly, AXO-resistant Enterobacterales counts also showed significant variations (p = 0.01), with fertilizer D exhibiting the highest density of resistant isolates. These findings indicate that fertilizer type significantly influences both bacterial abundance and the distribution of antimicrobial resistance (Table 2).

Dunn’s post hoc analysis identified fertilizer D as the main differentiating factor, showing a significantly higher microbial counts than the other treatments. Comparisons between group D and fertilizers B and C yielded the strongest statistical significance for both total Enterobacterales counts (p = 0.001) and AXO-resistant Enterobacterales (p = 0.0003). Additionally, fertilizer D differed significantly from fertilizer A (p = 0.016) and E (p = 0.013). In contrast, no significant differences were observed among groups A, B, C, and E (p > 0.05), indicating homogeneous distribution among these groups (

Table 3. Comparison of total Enterobacterales counts and AXO-resistant Enterobacterales across five fertilizers (Dunn’s post hoc test). Statistical significance is indicated by * (p < 0.05). AXO, ceftriaxone.

Comparison	Total Enterobacterales		AXO-Resistant Enterobacterales
	Delta Average Rank	p-Value	Delta Average Rank	p-Value
A vs. B	3.1	0.37	2.9	0.41
A vs. C	3.1	0.37	2.9	0.41
A vs. D	−8.4	* 0.016	−9.8	* 0.005
A vs. E	0.2	0.95	1	0.77
B vs. C	0	1	0	1
B vs. D	−11.5	* 0.001	−12.7	* 0.0003
B vs. E	−2.9	0.41	−1.9	0.58
C vs. D	−11.5	* 0.001	−12.7	* 0.0003
C vs. E	−2.9	0.41	−1.9	0.58
D vs. E	8.6	* 0.013	10.8	* 0.002

).

3.2. Identification of ESBL-Producing Isolates

A total of 19 lactose-fermenting isolates were recovered from 14 cultures showing growth under ceftriaxone selective conditions. Based on automated identification using the VITEK^® system, two species belonging to the order Enterobacterales were identified: E. coli (n = 13) and K. pneumoniae (n = 6). No additional lactose-fermenting morphotypes were detected. It is worth noting that a significant difference in the number of positive replicates was observed between industrial and artisanal fertilizers. Among the industrial formulations, bacterial growth was detected only in fertilizer FA (2/15 samples, 13.3%). In contrast, artisanal fertilizers showed heterogeneous results. Limited growth was observed in fertilizer FE (1/15 samples, 6.7%), a formulation composed primarily of humus, fruit peels, and eggshell residues. Conversely, fertilizer FD exhibited a substantially higher frequency of positive samples (11/15, 73.3%). This fertilizer consisted of a mixture of black soil, bovine manure, rice husks, and decomposing plant material. These findings suggest that the predominance of AXO-resistant isolates in fertilizer FD may be associated with the use of animal manure.

3.3. Antimicrobial Susceptibility Profiles and Phenotypic Detection of ESBL Production

In the double-disk synergy test, 84% (16/19) of the isolates tested positive for ESBL production, while 16% (3/19) yielded inconclusive results, specifically isolates FE5A-1, FD3A-1, and FD2B-1. Regarding antimicrobial susceptibility patterns, 100% of the isolates were resistant to ceftriaxone and cefotaxime, while 89.5% were resistant to ceftazidime (

Table 4. Antimicrobial susceptibility profiles of AXO-resistant Enterobacterales.

					MIC µg/mL
ID	Strain Identification	blaTEM	blaCTX-M	blaSHV	FOT	AXO	TAZ	FEP	FOX	P/T4	DORI	IMI	MERO	ETP	C/T	CZA
FA1B-1	E. coli				>4	>4	8	8	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4
FD1A-1	E. coli				>4	>4	32	8	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4
FD1B-1	E. coli				>4	>4	32	8	<8	4/4	<1	<1	<1	<0.5	2/4	<1/4
FD1C-1	E. coli				>4	>4	8	<2	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4
FD2A-2	E. coli				>4	>4	16	32	32	64/4	<1	<1	<1	<0.5	<1/4	<1/4
FD2B-1	E. coli				>4	>4	16	<2	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4
FD3A-2	E. coli				>4	>4	16	4	<8	8/4	<1	<1	<1	<0.5	2/4	<1/4
FD3B-1	E. coli				>4	>4	32	8	<8	8/4	<1	<1	<1	<0.5	<1/4	<1/4
FD3C-2	E. coli				>4	>4	16	4	<8	8/4	<1	<1	<1	<0.5	<1/4	<1/4
FD4A-2	E. coli				>4	>4	32	4	<8	4/4	<1	<1	<1	<0.5	2/4	<1/4
FD5A-2	E. coli				>4	>4	16	8	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4
FD5C-1	E. coli				>4	>4	32	8	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4
FE5A-1	E. coli				>4	>4	32	64	32	64/4	<1	<1	<1	<0.5	<1/4	<1/4
FA1C-1	K. pneumoniae				>4	>4	16	8	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4
FD2A-1	K. pneumoniae				>4	>4	16	<2	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4
FD3A-1	K. pneumoniae				>4	>4	32	64	32	32/4	<1	<1	<1	<0.5	<1/4	<1/4
FD3C-1	K. pneumoniae				>4	>4	16	8	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4
FD4A-1	K. pneumoniae				>4	>4	32	8	<8	8/4	<1	<1	<1	<0.5	<1/4	<1/4
FD5A-1	K. pneumoniae				>4	>4	16	4	<8	4/4	<1	<1	<1	<0.5	<1/4	<1/4

Abbreviations used in Table 4: FOT, cefotaxime; AXO, ceftriaxone; TAZ, ceftazidime; FEP, cefepime; FOX, cefoxitin; P/T4, piperacillin/tazobactam; DORI, doripenem; IMI, imipenem; MERO, meropenem; ETP, ertapenem; C/T, ceftolozane/tazobactam; CZA, ceftazidime/avibactam. Minimum inhibitory concentration (MIC) values are presented. Color coding indicates susceptibility interpretation according to CLSI criteria: green (susceptible), yellow (intermediate), and red (resistant). Dark blue: isolates carrying bla_TEM gene; orange: isolates carrying bla_CTX-M gene; dark green: isolates carrying bla_SHV gene; gray: isolates not carrying bla genes.

). Resistance to cefepime, cefoxitin, and piperacillin/tazobactam was also observed in several isolates. Quality control results were within CLSI acceptable ranges.

3.4. Detection and Distribution of bla_TEM, bla_SHV, and bla_{CTX-M.group 1} Genes Among Isolates

Endpoint PCR revealed that 100% of the isolates carried at least one of the screened genes. Among the 19 isolates, 68% co-expressed multiple resistance genes. Specifically, 42% (8/19) harbored bla_TEM and bla_CTX-M; 21% (4/19) harbored bla_SHV and bla_CTX-M; and 5% (1/19) simultaneously carried all three genes. Interestingly, PCR-based genotypic characterization revealed that three isolates with phenotypically inconclusive ESBL results carried bla genes, indicating false-negative outcomes in the synergy test. Such discrepancies are often attributed to the co-expression of ESBL and AmpC enzymes or hyperproduction of β-lactamases [28].

3.5. Genomic Characterization of Representative Isolates

Whole-genome sequencing (WGS) was performed for three representative isolates. The genome sizes of K. pneumoniae ranged from 5,486,733 to 5,544,155 bp, with an average GC content of approximately 57.2%. In contrast, the genome size of E. coli ranged from 4,919,848 to 4,942,912 bp, with a GC content of approximately 50.9%. Details of these studies are provided below:

(i) K. pneumoniae: As previously mentioned, two of the samples selected for the analysis were from artisanal organic fertilizer FD and industrial fertilizer FA. MLST analysis classified both isolates as sequence type 37 (ST37).

The β-lactam resistance phenotype was consistent with the genes identified. WGS confirmed that resistance to third-generation cephalosporins was mediated by the identified β-lactamase genes; however, bla_TEM was not amplified by conventional PCR. Additional resistance determinants included the colistin resistance mutation pmrB_R256G and genes conferring quinolone resistance, among others (Figure 2). Both isolates carried bla_CTX-M-55 gene.

Although none of the isolates exhibited a hypervirulent profile, they harbored several virulence factors associated with adherence, dissemination, and infection processes, including efflux pumps, siderophores, adhesins, porins, capsule, and type I and VI secretion systems (Figure 2).

(ii) E. coli: Based on WGS, the β-lactam resistance phenotype of E. coli was consistent with the detected genotype. MLST identified sequence type 224, which was also associated with the presence of bla_CTX-M-55. Additional resistance genes conferring resistance to colistin, fosfomycin, and quinolones were also identified. As for the virulome, a greater proportion of adhesion-associated genes were observed, followed by those for siderophores, invasion, and porins (Figure 3).

4. Discussion

Organic fertilizers are widely recognized for their agronomic and environmental benefits, as they contain organic and mineral components free of synthetic additives. However, they have not been widely acknowledged as potential reservoirs and dissemination sources of ESBL-producing bacteria. This concern is particularly relevant because current regulations, both globally and in Colombia, allow the use of animal- or plant-based raw materials, as well as biosolids and residues from domestic or agro-industrial wastewater, without considering routine antimicrobial use in these contexts [29,30]. In this context, the present study provides the first evidence from Colombia demonstrating the occurrence of ESBL-producing Enterobacterales in commercially available organic fertilizers.

Notably, AXO-resistant Enterobacterales were detected in 18.6% of the analyzed samples, further supporting the role of organic fertilizers as environmental reservoirs of antibiotic-resistant bacteria. Although the present study was not designed as a population-based survey, the observed resistance frequencies of ceftriaxone and ceftazidime were comparable to those reported in clinical isolates by the Colombian National Institute of Health in 2020 [31]. Therefore, this finding suggests that resistance mechanisms commonly associated with healthcare settings may also be present in agricultural environments, highlighting the need to integrate One Health approaches into national policies, as already implemented in Europe [32,33].

Multiple factors may contribute to the presence of resistant bacteria in fertilizers [34,35]. As previously pointed out, current regulations permit the incorporation of materials that may originate from environments heavily contaminated with antibiotics [4]. Antibiotics administered for therapeutic purposes or as growth promoters in livestock are often excreted in active or partially metabolized forms through feces and urine. When these residues are incorporated into manure-based fertilizers, they may accumulate in soils and interact with native microbial communities, potentially facilitating the persistence and horizontal transfer of antimicrobial resistance genes (ARGs) [36]. Consequently, resistant bacteria and resistance determinants may disseminate through environmental pathways such as agricultural runoff, surface waters, and indirect exposure along the food production chain [37].

Our statistical analysis confirms that the type of fertilizer significantly influences the colony count (p = 0.001) (Table 2). In particular, the manure-amended formulation (fertilizer D) showed the highest mean colony counts, forming an independent subset, whereas other fertilizers exhibited complete inhibition under the tested conditions. These findings are consistent with previous studies reporting that manure-amended soils can harbor diverse ARGs associated with multiple antimicrobial classes [34,38]. Although β-lactams are typically introduced through manure, subsequent microbial interactions in the soil environment can promote the emergence and proliferation of additional ARGs [36]. Notably, these genes have been detected more abundantly in leaf endophytes than in roots, indicating that manure-derived ARGs can accumulate in aboveground plant tissues, increasing their potential dissemination within agricultural ecosystems [39].

While the transfer of antimicrobial-resistant bacteria from soil to plant products remains complex and not fully understood, several factors, including agricultural practices, environmental conditions, and post-harvest handling may influence this process [34,36]. Therefore, although our study does not directly assess transmission to food products, the presence of ESBL-producing Enterobacterales in fertilizers highlights a potential pathway for indirect dissemination within agroecosystems [37,38].

In Colombia, this situation is compounded by unregulated access to veterinary antibiotics. The Colombian Agricultural Institute currently lists >60 veterinary products whose main active ingredient is a third-generation cephalosporin (ceftiofur) commonly used in cattle and swine, which are primary sources of manure used for fertilizer [11]. Given the structural similarity between ceftiofur and clinically relevant cephalosporins such as ceftriaxone and cefotaxime, shared selective pressures across agricultural and clinical sectors may contribute to the maintenance and spread of resistance determinants [40].

At the molecular level, the predominance of bla_CTX-M observed in this study aligns with global reports describing the successful dissemination of this gene family across human, animal, and environmental compartments. Studies of E. coli isolates have demonstrated an alarming expansion across different types of samples analyzed, including broiler chicken [41], water reservoirs in wastewater treatment plants, and recreational [42], swine manure [38], among others.

K. pneumoniae ST37 has previously been reported in clinical settings as a producer of different β-lactamases, such as CTX-M-15 and KPC [42,43]. It has also been associated with resistance of phenotypes to ceftazidime/avibactam and colistin [44]. Reports have also documented the presence of ST37 in environmental samples [45] from around dairy cows in China [46], marine environments in Norway [47], wastewater treatment plants in Europe [48], companion animals in China [49], and surface waters [50]. In Colombia, ST37 has been reported in humans [51]. However, to date, there are no reports of its presence in environmental sources.

Similarly, E. coli ST224 has been mainly associated with environmental and animal sources. It has been reported in pet owners previously infected or colonized with this ST [52], as well as in urinary tract infection (UTIs) [53] and in cases of colistin and carbapenem resistance [54]. Although phylogroup B1 was initially defined as a commensal strain [55], different studies have linked it to high-risk clones, such as ST131 carriers of CTX-M [56].

The detection of the CTX-M-55 variant in both species is of particular importance. A systematic review by Yu et al. [57] revealed CTX-M-55 as a dominant variant in Asia. This variant has been frequently detected in E. coli isolated from humans [52], as well as in eggs [58], poultry [59], and even in herbs such as coriander [60]. Its detection in industrial fertilizers in the present study indicates the regional dissemination of globally relevant resistance determinants.

Most studies have focused on the emergence and spread of ESBL genes in hospital environments, wastewater treatment plants, aquatic ecosystems, and animal-derived food products [61,62,63]. ESBL-producing E. coli can be disseminated from animal farms to the surrounding environments, including rural water reservoirs [53,63]. However, the role of animal-derived composting materials in disseminating ESBL genes remains understudied. Said et al. [14] identified ESBL-producing isolates in 7.31% (3/41) of agricultural soil samples and in 21.93% (25/114) of positive wastewater and surface water samples. Similarly, Gao et al. [38] used Enterobacterial repetitive intergenic consensus–PCR typing to demonstrate that composted soil isolates originated from pig feces used in the composting process.

Although composting is often proposed as a mitigation strategy for reducing the microbial and chemical risks associated with manure reuse in agriculture, its efficiency in removing ARGs is highly variable, and certain resistance determinants can persist despite the treatment [36,53]. Most previous investigations have focused on resistance genes related to tetracyclines, macrolides, or sulfonamides, whereas data on β-lactamase genes in composted agricultural inputs remain scarce.

Because current Colombian technical standards do not include routine screening for fertilizers or food production, these products may contribute to the environmental dissemination of resistance. To the best of our knowledge, this study provides the first evidence of ESBL-producing Enterobacterales in organic fertilizers marketed in Colombia. Furthermore, these findings underscore the importance of incorporating microbiological surveillance of agricultural inputs into national antimicrobial resistance mitigation strategies under a comprehensive One Health approach.

5. Study Limitations: Brief Discussions

This study has several limitations that should be acknowledged. First, fertilizer types were selected based on market availability rather than probabilistic sampling, which may limit the generalizability of the findings. Second, bacterial enumeration was performed following enrichment steps, and therefore CFU values should be interpreted as semi-quantitative estimates of recoverable Enterobacterales rather than absolute bacterial loads. Additionally, only a limited number of isolates were subjected to whole-genome sequencing, which restricts broader conclusions regarding the genomic diversity and dissemination patterns of resistance determinants. Finally, this investigation focused on microbiological characterization and did not evaluate other parameters.

Despite these limitations, this study appears to be the first report from Colombia describing ESBL-producing bacteria in organic fertilizers. These findings provide novel baseline evidence and support the need for integrated surveillance approaches under the One Health framework.

6. Conclusions

This study demonstrates the presence of extended-spectrum β-lactamase (ESBL)-producing Enterobacterales in both industrial and artisanal organic fertilizers commercially available in Southwestern Colombia. These findings suggest that organic fertilizers may act as environmental reservoirs and potential dissemination pathways for antimicrobial-resistant bacteria within agricultural ecosystems. For example, significant differences in recoverable Enterobacterales counts among fertilizer types suggest variability in microbiological quality likely influenced by production practices and raw material composition. Furthermore, the detection of clinically relevant resistance genes, together with the identification of multidrug-resistant isolates and genomic characterization of representative strains, reinforces concerns regarding the environmental circulation of antimicrobial resistance determinants.

From a One Health perspective, the use of untreated or poorly controlled organic fertilizers may contribute to the introduction and persistence of resistant bacteria in soils, crops, and potentially the food chain. Although direct transmission to humans or food products was not evaluated in this study, the findings reported in this contribution support the need for strengthened microbiological surveillance and regulatory oversight of organic fertilizer production and commercialization.

Based on the results reported in this contribution, several important aspects still require further investigation. Future research should focus on the survival dynamics, transmission pathways, and public health implications of antimicrobial-resistant bacteria associated with agricultural inputs to better inform risk mitigation strategies.