Prevalence of antimicrobial resistance in equine-associated Pseudomonas aeruginosa: a systematic review and meta-analysis

Abstract

The escalating issue of antimicrobial resistance (AMR) in Pseudomonas aeruginosa (P. aeruginosa), particularly in the equine context, poses a significant threat to veterinary and public health. This systematic review and meta-analysis aimed to assess the current landscape of AMR in P. aeruginosa strains isolated from equine sources. Adhering to the PRISMA guidelines, a comprehensive literature search was conducted, and data from eligible studies were extracted and synthesized. The review included 10 articles reporting on 1624 P. aeruginosa isolates. High resistance rates were found for several antimicrobial classes, with particular concern for increasing resistance to imipenem, amikacin, and ceftiofur. The study also identifies significant temporal trends and regional variations in resistance patterns. Asia reported the highest average resistance rates, suggesting potential misuse of antimicrobials in the region. The most effective antimicrobials were found to be aztreonam, fosfomycin, ciprofloxacin, and ceftazidime. However, the generalizability of these findings is limited by the small number of eligible studies (n = 10) and the scarcity of isolates for certain antimicrobial classes. The findings of this review highlight the critical need for innovative strategies, stringent antimicrobial stewardship, and comprehensive risk assessment to combat the escalating threat of AMR in equine-associated P. aeruginosa and preserve the efficacy of our existing antimicrobial agents.

Introduction

Pseudomonas aeruginosa (P. aeruginosa), is a common opportunistic pathogen known for its extensive antimicrobial resistance (AMR) and is a leading cause of chronic or acute nosocomial infections in humans [1–3]. In equine veterinary medicine, infections with P. aeruginosa are relatively rare [4], commonly seen as secondary infections. P. aeruginosa infection can be difficult to treat effectively, with relatively poor recovery and prognosis. In equine animals, including horses and donkeys P. aeruginosa has been reported causing endometritis and infertility [5–7]. It can also cause secondary polymicrobial infections in the reproductive, respiratory, gastrointestinal, ocular, dermal, and pharyngeal sacs after antimicrobial treatment [1, 8–10].

Antimicrobial treatment of P. aeruginosa infections is challenging due to its intrinsic resistance to many types of antimicrobials, combined with the ability to acquire new resistances, making resistance and multidrug resistance very common [11]. These intrinsic resistances are the results of genetic mutations; some mutations may render certain antimicrobials completely ineffective, while others may lead to a gradual accumulation of resistance to most antimicrobials [12–14], without immediately manifesting noticeable clinical significance. Resistance can be caused by various mechanisms, including efflux pumps on the bacterial cell membrane, genetic mutations in antimicrobial target sites, production of enzymes such as β-lactamases that destroy the active parts of antimicrobials, and other mechanisms [15]. Additionally, its adherence to tissues, intracellular accumulation, and biofilm formation further increases the difficulty of antimicrobial treatment [16, 17]. The presence and extent of intrinsic resistance of P. aeruginosa to specific classes of antimicrobials are associated with different strains [18]. However, there is a scarcity of investigations concerning these aspects in P. aeruginosa strains derived from equine source. Furthermore, in equine veterinary medicine, the limited variety of legally available antimicrobials means that the long-term use of the same antimicrobials may exacerbate resistance [1]. The increasing prevalence of AMR in equine P. aeruginosa isolates necessitates a reassessment of our methods for managing bacterial infections in horses.

In common clinical infectious diseases of horses and donkeys, such as trauma [19], reproductive tract infections [8, 9], corneal ulcers [20, 21], and infectious synovitis [22], particularly when the affected areas are in extensive contact with the environment [5, 23], P. aeruginosa often forms complex polymicrobial infections alongside other types of pathogens. Sometimes, horses may not exhibit clinical symptoms of infection. However, carriers (mares and stallions) can still cause transmission [24]. This makes treatment and recovery more difficult and the prognosis more pessimistic. Especially in the case of corneal infections [25, 26] and infectious synovitis [27, 28], if treatment is not timely or the therapeutic effect is poor, it may cause permanent damage to the cornea and synovial membrane. Infections of the reproductive tract may cause infertility in female livestock [4, 8]. These issues can affect the athletic performance [6] and reproductive performance [8] of horses, leading to significant loss of economic benefits and posing a great threat to their quality of life and welfare.

Current literature lacks a comprehensive assessment of P. aeruginosa AMR patterns in equine sources around the world. In this context, the purpose of this review is to study the complexity of AMR of P. aeruginosa in equine animals. By examining the latest research, our goal is to provide a comprehensive overview of the current state of AMR in P. aeruginosa in equine animals and discuss potential strategies to mitigate its impact on health. This issue extends beyond veterinary medicine, posing an urgent global public health challenge.

Materials and methods

Study design

This systematic review and meta-analysis aimed to assess the current landscape of AMR in P. aeruginosa strains isolated from equine sources. The study was conducted following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines to ensure a rigorous and transparent approach.

Data sources and search strategy

A comprehensive literature search was conducted on 25 August 2024 in PubMed, Scopus and Web of Science from January 1990 to July 2024. The search strategy included a combination of MeSH terms and keywords related to equine medicine, P. aeruginosa, and AMR. The search syntax was tailored to each database and included terms such as “equine”, “P. aeruginosa”, “antibiotic resistance”, and “antimicrobial agents”. The detailed search string is provided in Table 1.

Table 1.

Search terms utilized for the systematic review construction

Antimicrobials	Population	Focus	Pathogen
antimicrobial* OR anti-microbial* OR antibiotic* OR anti-biotic* OR bacteriocide* OR bacteriostat* OR anti-infective OR microbicide OR antiseptic OR antibacterial	AND (equine* OR horse* OR Equidae OR Equu* OR equestrian* OR foal* OR pony OR donkey* OR mule* OR burro* OR ass OR asses[tw] OR jackass* OR moke* OR jenny OR jennies OR jennet* OR jack* OR hinnies OR cuddies OR asinine OR mulish OR zebra* OR quagga OR “stripped horse*”)	AND (resis* OR risk OR “risk factor” OR driver OR resistance OR resistant OR efficacy OR effectiveness)	AND (“Pseudomonas aeruginosa” OR “P. aeruginosa”)

‘AND’ signifies that the terms were merged in the search

Inclusion and exclusion criteria

Studies were included if they reported on the prevalence or patterns of AMR in P. aeruginosa strains derived from horses or other equine species. Studies were excluded if they focused on other bacterial species, non-equine sources, or did not provide sufficient data on AMR patterns. Additionally, studies not published in English or not original research articles (e.g., reviews, editorials, and case reports) were excluded.

Data extraction and quality assessment

Two independent reviewers extracted data from eligible studies using a standardized extraction form. The extracted data included study characteristics (e.g. authors, year of publication, country), sample size, antimicrobial agents tested, and resistance rates. The quality of the studies was assessed using a modified Newcastle-Ottawa Scale, focusing on the selection of studies, comparability, and exposure.

Statistical analysis

Data analysis was conducted using appropriate statistical procedures in the SPSS software package. Meta-analytical techniques were employed to pool the data on AMR rates across studies. The random-effects model was used to account for the heterogeneity between studies. The I² statistic was used to measure the degree of heterogeneity, and a P-value < 0.05 was considered statistically significant. Subgroup analyses were conducted based on geographical location, study period, and the type of antimicrobials tested. When analyzing the differences in the resistance rates (RR) between different groups, if the expected frequency was less than 5, Fisher’s exact test was employed to assess the significance, otherwise the Chi-square test was used to examine the significance.

Ethical considerations

This study involved the analysis of existing literature and did not require ethical approval. All data were analyzed anonymously, and no personal information from individual studies was included in this review.

Results

Summary of articles

In this review, we meticulously evaluated 601 documents: 569 key articles were discovered via databases such as PubMed and Web of Science, an additional 28 reports were extracted from grey literature searches, and a further 4 studies were found by scrutinizing the bibliographies of the principal articles. We excluded 1 article not in English. Following the initial review of titles and abstracts, we conducted a full-text analysis of 23 articles, excluding 13 that did not align with the population, intervention, comparison, outcomes of an article, and study design framework [29]. This process resulted in 10 articles [4, 6, 7, 20, 27, 30–34] being included for data synthesis (Fig. 1).

Fig. 1.

Flowchart of literature retrieval and criteria used to select documents for the systematic review

Meta-analysis results

The RR of 1624 bacterial isolates from the 10 included studies to various antimicrobial agents were summarized (Table 2; Fig. 2). The degree of heterogeneity is shown by I² statistic listed in Table 2. In this study, AMR was defined according to the criteria reported in the original articles. Where available, susceptibility interpretations were based on internationally recognized guidelines, primarily those issued by the Clinical and Laboratory Standards Institute (CLSI) or the European Committee on Antimicrobial Susceptibility Testing (EUCAST). For studies that did not explicitly state the guideline used, the resistance status reported by the authors was accepted as defined. This approach ensured the inclusion of the broadest range of studies, although it may introduce some variability due to differing interpretative standards.

Table 2.

Characteristics of the included studies

Class of antimicrobials	Antimicrobials	Effect Size									Heterogeneity
		Number of studies	Number of resistant isolates	Number of total isolates	Proportion	SE	Low Confidence interval	High Confidence interval	Z value	Z value p-value	I2
Penicilins	Piperacillin	1	12	135	0.09	0.02	0.04	0.14	3.63	0.00
	Piperacillin-Tazobactam	1	7	135	0.05	0.02	0.01	0.09	2.72	0.01
	Ticarcillin	2	9	139	0.07	0.02	0.02	0.11	3.11	0.00	75.00
	Ticarcillin-Clavulanic Acid	2	11	139	0.08	0.02	0.03	0.12	3.45	0.00	75.00
	Amoxicillin-Clavulanic-Acid	2	60	77	0.78	0.05	0.69	0.87	16.49	0.00	0.00
	Ampicillin	2	10	12	0.85	0.10	0.64	1.00	8.14	0.00	0.00
	Penicillin	2	11	12	0.93	0.08	0.78	1.00	12.31	0.00	75.00
	Oxacillin	1	4	4	0.99	0.05	0.89	1.00	19.90	0.00
	Amoxicillin	1	146	146	0.99	0.01	0.97	1.00	120.22	0.00
Carbapenems	Imipenem	3	11	140	0.08	0.02	0.03	0.12	3.43	0.00	0.00
	Meropenem	2	14	142	0.10	0.03	0.05	0.15	3.94	0.00	75.00
	Ertapenem	1	1	1	0.99	0.10	0.79	1.00	9.95	0.00
Monobactams	Aztreonam	1	1	135	0.01	0.01	0.00	0.02	0.98	0.33
Cephalosporins	Cefquinome	4	587	1524	0.39	0.01	0.36	0.41	30.90	0.00	24.67
	Ceftiofur	5	189	363	0.52	0.03	0.47	0.57	19.85	0.00	0.00
	Cefepime	2	3	142	0.02	0.01	0.00	0.04	1.74	0.08	75.00
	Ceftazidime	1	0	135	0.01	0.01	0.00	0.03	1.17	0.24
	Ceftolozane-Tazobactam	1	0	135	0.01	0.01	0.00	0.03	1.17	0.24
	Cefazolin	1	7	7	0.99	0.04	0.92	1.00	26.32	0.00
	Cefoxitin	1	7	7	0.99	0.04	0.92	1.00	26.32	0.00
	Cephalexin	1	8	8	0.99	0.04	0.92	1.00	28.14	0.00
	Cefadroxil	1	10	10	0.99	0.03	0.93	1.00	31.46	0.00
Phosphonic acids	Fosfomycin	1	1	135	0.01	0.01	0.00	0.02	0.98	0.33
Aminoglycosides	Amikacin	7	143	1259	0.11	0.01	0.10	0.13	12.69	0.00	0.00
	Gentamicin	9	469	1607	0.29	0.01	0.27	0.31	25.74	0.00	0.00
	Netilmicin	1	48	135	0.36	0.04	0.27	0.44	8.62	0.00
	Tobramycin	1	14	135	0.10	0.03	0.05	0.16	3.96	0.00
	Kanamycin	1	7	7	0.99	0.04	0.92	1.00	26.32	0.00
	Neomycin	1	6	8	0.71	0.16	0.40	1.00	4.43	0.00
Fluoroquinolones	Enrofloxacin	5	116	360	0.32	0.02	0.27	0.37	13.08	0.00	0.00
	Marbofloxacin	6	110	1533	0.07	0.01	0.06	0.08	10.89	0.00	0.00
	Ciprofloxacin	2	4	143	0.03	0.01	0.00	0.06	2.05	0.04	0.00
	Levofloxacin	1	12	135	0.09	0.02	0.04	0.14	3.63	0.00
Tetracyclines	Doxycycline	3	53	82	0.65	0.05	0.54	0.75	12.24	0.00	0.00
	Tetracycline	3	156	158	0.99	0.01	0.97	1.00	111.01	0.00	80.00
	Oxytetracycline	1	61	70	0.87	0.04	0.79	0.95	21.78	0.00
Sulfonamides	Trimethoprim-Sulfamethoxazole	5	209	235	0.89	0.02	0.85	0.93	43.47	0.00	72.36
Rifamycins	Rifampicin	2	10	11	0.93	0.08	0.77	1.00	11.84	0.00	0.00
Amphenicols	Chloramphenicol	3	67	82	0.82	0.04	0.73	0.90	19.04	0.00	0.00
Macrolides	Azithromycin	1	4	4	0.99	0.05	0.89	1.00	19.90	0.00
	Erythromycin	1	4	4	0.99	0.05	0.89	1.00	19.90	0.00
Lincosamides	Clindamycin	1	8	8	0.99	0.04	0.92	1.00	28.14	0.00
Peptide Antibiotics	Bacitracin	1	8	8	0.99	0.04	0.92	1.00	28.14	0.00
	Polymyxin B	1	7	8	0.86	0.12	0.62	1.00	7.01	0.00

Fig. 2.

The prevalence of resistance for each antimicrobial

Resistance to penicillins class

Piperacillin and piperacillin-tazobactam

A study determined the susceptibility to piperacillin and piperacillin-tazobactam, which included 135 isolates of P. aeruginosa. Twelve and seven P. aeruginosa isolates were found to be resistant to piperacillin and piperacillin-tazobactam, respectively. The RR for piperacillin was 9% (95% CI 4%−14%), and for piperacillin-tazobactam it was 5% (95% CI 1%−9%), with no significant difference between the two (P > 0.05).

Temporal trends could not be assessed as only a single dataset was available for the period 1996–2020, which reported the relative risk as indicated in Fig. 3.

Fig. 3.

The heat map of the prevalence of resistance of antimicrobials during the timeNote: a or b: in the same row, the same lowercase letters indicate that there is no significant difference between the two values

In the continental subgroup analysis, only one set of data from Europe was available (Table 2), limiting insights into variations across different continents.

Ticarcillin and ticarcillin-clavulanic acid

In two studies, the susceptibility to ticarcillin and ticarcillin-clavulanic acid was determined, encompassing a total of 139 isolates. Nine and one isolates were found to be resistant to ticarcillin and ticarcillin-clavulanic acid, respectively. The RR for ticarcillin was 7% (95% CI 2%−11%), and for ticarcillin-clavulanic acid it was 8% (95% CI 3%−12%), with no significant difference between them (P > 0.05). There was considerable heterogeneity among the literature (I² ≥ 75%).

Only two datasets, covering the periods 1996–2020 and 2010–2015, provided information on the RR (Fig. 3), making it difficult to determine trends or changes over time.

The continental subgroup analysis revealed that the RR reported in Europe was significantly higher (P < 0.05) than that in North America, with ticarcillin and ticarcillin-clavulanic acid both at 7% and 8%, respectively, while in North America both were at 0 (Fig. 4).

Fig. 4.

The heat map of the prevalence of resistance in different continentsNote: a, b, c, or d: in the same row, the same lowercase letters indicate that there is no significant difference between the two values

Amoxicillin

One study determined the susceptibility to amoxicillin, including 146 isolates, all of which were found to be resistant. The RR was 100% (95% CI 97%−100%). Only one data set from 2018 to 2022 in Europe was available (Figs. 3 and 4).

Amoxicillin-clavulanic acid

In two studies, the susceptibility to amoxicillin-clavulanic acid was determined, totaling 77 isolates. Sixty isolates were found to be resistant, with an RR of 78% (95% CI 69%−87%), which was significantly different from that of amoxicillin (P < 0.01). The heterogeneity among the literature was minimal (I² < 25%).

The limitation to discern temporal trends is imposed by the constraint that only two datasets, encompassing the periods 2010–2020 and 2010–2015, report the relative risk as indicated in Fig. 3.

In the continental subgroup analysis, only one set of data from Asia was available (Fig. 4), limiting an understanding of the variations across different continents.

Ampicillin

In two studies, the susceptibility to ampicillin was determined, totaling 12 isolates. Ten were found to be resistant, with an RR of 7% (95% CI 2%−11%). The heterogeneity among the literature was minimal (I² < 25%).

Only two datasets, sampled 2010–2020 and 2010–2015, provided information on the RR (Fig. 3), making it difficult to determine trends or changes over time.

The continental subgroup analysis showed no significant differences between continents (P > 0.05), with North America reporting a lower RR at 75%, and Oceania at 88% (Fig. 4).

Penicillin

In two studies, the susceptibility to penicillin was determined, with a total of 12 isolates examined. Eleven strains were found to be resistant, with an RR of 93% (95% CI 78%−100%). There was considerable heterogeneity among the literature (I² ≥ 75%).

The limitation to discern temporal trends is imposed by the constraint that only two datasets, encompassing the periods 2010–2020 and 2010–2015, report the relative risk as indicated in Fig. 3.

The continental subgroup analysis showed no significant differences between continents (P > 0.05), with North America reporting a higher RR at 100%, and Oceania at 88% (Fig. 4).

Oxacillin

One study determined the susceptibility to oxacillin, including four isolates, all of which were found to be resistant, with an RR of 100% (95% CI 89%−100%). Only one data set from 2010 to 2015 in North America was available (Figs. 3 and 4).

Resistance to carbapenems class

Imipenem

In three studies, the susceptibility to Imipenem was determined with 140 isolates in total. Eleven were found to be resistant, with an RR of 8% (95% CI 3%−12%). The heterogeneity among the literature was minimal (I² < 25%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a significant increase in RR (20%−100%, P < 0.01) (Fig. 3).

There was a highly significant difference between continents (P < 0.01), with North America reporting the highest RR at 100%, and Europe at 7% (Fig. 4).

Meropenem

In two studies, the susceptibility to meropenem was determined, involving a total 142 isolates. Fourteen were found to be resistant, with an RR of 10% (95% CI 5%−15%). There was considerable heterogeneity among the literature (I² ≥ 75%).

Only two datasets, covering the periods 1996–2020 and 2018–2021, provided information on the RR (Fig. 3), making it difficult to determine trends or changes over time.

The continental subgroup analysis showed a highly significant difference between continents (P < 0.01), with Asia reporting a higher RR at 100%, and Europe at 5% (Fig. 4).

Ertapenem

One study determined the susceptibility to ertapenem, including one isolate which was found to be resistant, with an RR of 100% (95% CI 79%−100%). Only one data set from 2023 in Europe was available (Figs. 3 and 4).

Resistance to aztreonam of monobactams class

One study determined the susceptibility to aztreonam, including 135 isolates, one of which was found to be resistant, with an RR of 1% (95% CI 0–2%). Only one data set from 1996 to 2020 in Europe was available (Figs. 3 and 4).

Resistance to cephalosporins class

Cefquinome

In three studies, the susceptibility to cefquinome was determined, involving 1524 isolates. Five hundred eighty-seven were found to be resistant, with an RR of 39% (95% CI 36%−41%). The heterogeneity among the literature was minimal (I² < 25%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a significant increase in RR (39%−48%, P < 0.01) (Fig. 3).

In the continental subgroup analysis, only data from Europe was available (Fig. 4), which did not allow for an understanding of the variations across different continents.

Ceftiofur

In five studies, the susceptibility to ceftiofur was determined, totalling 363 isolates. One hundred eighty-nine were found to be resistant, with an RR of 52% (95% CI 47%−57%). The heterogeneity among the literature was minimal (I² < 25%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a significant increase in RR (20%−90%, P < 0.01) (Fig. 3).

The continental subgroup analysis showed a significant difference between continents, with North America reporting the lowest RR at 25%, and a significant increase in Asia to 69% (P < 0.01), Oceania and Europe also significantly increased (P < 0.01), to 88% and 99%, respectively, but with no significant difference between the two (Fig. 4).

Cefepime

In two studies, the susceptibility to cefepime was determined, involving a total of 142 isolates. Three were found to be resistant, with an RR of 2% (95% CI 0–4%). There was considerable heterogeneity among the literature (I² ≥ 75%).

In the time subgroup analysis, only two sets of data covering the periods 1996–2020 and 2018–2021 were available (Fig. 4), limiting an understanding of the variations between different periods.

The continental subgroup analysis showed a significant difference between continents (P < 0.05), with Europe reporting a slightly higher RR at 2%, and Asia at 0 (Fig. 4).

Ceftazidime and ceftolozane-tazobactam

One study determined the susceptibility to ceftazidime and ceftolozane-tazobactam, including 135 isolates, none of which were found to be resistant, with an RR of 0 (95% CI 0–3%). Only one data set from 1996 to 2020 in Europe was available (Figs. 3 and 4).

Cefazolin and cefoxitin

One study determined the susceptibility to cefazolin and cefoxitin, including seven isolates, all of which were found to be resistant, with an RR of 99% (95% CI 92%−100%). Only one data set from 2018 to 2021 in Asia was available (Figs. 3 and 4).

Cephalexin

One study determined the susceptibility to cephalexin, including eight isolates, all of which were found to be resistant, with an RR of 99% (95% CI 92%−100%). Only one data set from 2010 to 2020 in Oceania was available (Figs. 3 and 4).

Cefadroxil

One study determined the susceptibility to cefadroxil, including ten isolates, all of which were found to be resistant, with an RR of 99% (95% CI 93%−100%). Only one data set from 1985 in North America was available (Figs. 3 and 4).

Resistance to fosfomycin of phosphonic acids class

One study determined the susceptibility to fosfomycin, including 135 isolates, one of which was found to be resistant, with an RR of 1% (95% CI 0–2%). Only one data set from 1996 to 2020 in Europe was available (Figs. 3 and 4).

Resistance to aminoglycosides class

Amikacin

In five studies, the susceptibility to amikacin was determined, totaling 1259 isolates. One hundred forty-three were found to be resistant, with an RR of 11% (95% CI 10%−13%). The heterogeneity among the literature was minimal (I² < 25%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a significant increase in RR (7%−86%, P < 0.01) (Fig. 3).

The continental subgroup analysis showed that RR varied from lowest to highest as follows: Oceania, Europe, North America, and Asia, with 0%, 6%, 25%, and 87% respectively. Except for the non-significant difference between Europe and both Oceania and North America, all other pairwise comparisons were significantly different (P < 0.05).

Gentamicin

In five studies, the susceptibility to gentamicin was determined, with a total of 1607 isolates involved. Four hundred sixty-nine were found to be resistant, with an RR of 29% (95% CI 27%−31%). The heterogeneity among the literature was minimal (I² < 25%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a significant decrease in RR (31%−28%, P < 0.01) (Fig. 3).

The continental subgroup analysis showed that RR varied from lowest to highest as follows: Europe, Oceania, North America, and Asia, with 26%, 38%, 50%, and 88% respectively, and all pairwise comparisons were significantly different (P < 0.05).

Netilmicin and tobramycin

One study determined the susceptibility to netilmicin and tobramycin, including 135 isolates, of which 48 and 14 were found to be resistant, with RRs of 36% (95% CI 27%−44%) and 10% (95% CI 5%−16%), respectively. Only one data set from 1996 to 2020 in Europe was available (Figs. 3 and 4).

Kanamycin

One study determined the susceptibility to kanamycin, including seven isolates, all of which were found to be resistant, with an RR of 99% (95% CI 92%−100%). Only one data set from 2018 to 2021 in Asia was available (Figs. 3 and 4).

Neomycin

One study determined the susceptibility to neomycin, including eight isolates, six of which were found to be resistant, with an RR of 71% (95% CI 40%−100%). Only one data set from 2010 to 2020 in Oceania was available (Figs. 3 and 4).

Resistance to fluoroquinolones class

Enrofloxacin

In five studies, the susceptibility to enrofloxacin was determined, totaling 360 isolates. One hundred sixteen were found to be resistant, with an RR of 32% (95% CI 27%−37%). The heterogeneity among the literature was minimal (I² < 25%).

The overlapping time periods of the five studies made it difficult to discern any temporal trends.

The continental subgroup analysis showed that RR varied from lowest to highest as follows: Europe, Oceania, and Asia, with 20%, 50%, and 81% respectively, and all pairwise comparisons were significantly different (P < 0.05).

Marbofloxacin

In three studies, the susceptibility to marbofloxacin was determined, including 1533 isolates. One hundred ten were found to be resistant, with an RR of 7% (95% CI 6%−8%). The heterogeneity among the literature was minimal (I² < 25%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a significant decrease in RR (31%−28%, P < 0.01) (Fig. 3).

The continental subgroup analysis showed a significant difference between continents (P < 0.00), with Europe reporting a lower RR at 7%, and Oceania at 50% (Fig. 4).

Ciprofloxacin

In two studies, the susceptibility to ciprofloxacin was determined, involving 143 isolates. Four were found to be resistant, with an RR of 3% (95% CI 0–6%). The heterogeneity among the literature was minimal (I² < 25%).

The overlapping time periods of the two studies made it impossible to discern any temporal trends.

The continental subgroup analysis showed a significant difference between continents (P < 0.01), with Europe reporting a lower RR at 2%, and Oceania at 13% (Fig. 4).

Levofloxacin

One study determined the susceptibility to levofloxacin, including 135 isolates, twelve of which were found to be resistant, with an RR of 9% (95% CI 4%−14%). Only one data set from 1996 to 2020 in Europe was available (Figs. 3 and 4).

Resistance to tetracyclines class

Doxycycline

In three studies, the susceptibility to doxycycline was determined, totaling 82 isolates. Fifty-three were found to be resistant, with an RR of 65% (95% CI 54%−75%). The heterogeneity among the literature was minimal (I² < 25%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a significant decrease in RR (100%−64%, P < 0.01) (Fig. 3).

The continental subgroup analysis showed that RR varied from lowest to highest as follows: Oceania, Asia, and North America, with 50%, 64%, and 100% respectively, but differences between them were not significant (P > 0.05).

Tetracycline

In 3 studies, the susceptibility to tetracycline was determined and included 158 isolates. One hundred fifty-six were found to be resistant, with an RR of 99% (95% CI 97%−100%). There was considerable heterogeneity among the literature (I² ≤ 75%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated no change in RR, which remained at 100% (Fig. 3).

The continental subgroup analysis showed no significant differences between continents (P > 0.05), with North America reporting a higher RR at 100%, and Oceania at 75% (Fig. 4).

Oxytetracycline

One study determined the susceptibility to oxytetracycline, including seventy isolates, sixty-one of which were found to be resistant, with an RR of 87% (95% CI 79%−95%). Only one data set from 2016 to 2021 in Asia was available (Figs. 3 and 4).

Resistance to trimethoprim-sulfamethoxazole of sulfonamides class

In three studies, the susceptibility to trimethoprim-sulfamethoxazole was determined, involving 235 isolates. Two hundred nine were found to be resistant, with an RR of 89% (95% CI 85%−93%). The heterogeneity among the literature was moderate (50% < I² < 75%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a significant decrease in RR (100%−88%, P < 0.01) (Fig. 3).

The continental subgroup analysis showed that RR varied from lowest to highest as follows: Asia, Europe, Oceania, and North America, with 70%, 98%, 100%, and 100% respectively, and Asia significantly differed from other regions (P < 0.01).

Resistance to rifampicin of rifamycins class

In two studies, the susceptibility to rifampicin was determined, totaling eleven isolates. Ten were found to be resistant, with an RR of 93% (95% CI 77%−100%). The heterogeneity among the literature was minimal (I² < 25%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a non-significant increase in RR (80%−100%, P > 0.05) (Fig. 3).

The continental subgroup analysis showed that Asia reported a higher RR at 100%, and North America at 75% (Fig. 4), with no significant difference between them.

Resistance to chloramphenicol of amphenicols class

In three studies, the susceptibility to chloramphenicol was determined, totaling 82 isolates. Sixty-seven were found to be resistant, with an RR of 82% (95% CI 73%−90%). The heterogeneity among the literature was minimal (I² < 25%).

The year subgroup analysis between two groups (2006–2016 and 2016–2023) indicated a significant decrease in RR (100%−80%, P < 0.01) (Fig. 3).

The continental subgroup analysis showed that RR varied from lowest to highest as follows: Asia, Oceania, and North America, with 80%, 88%, and 100% respectively, and all pairwise comparisons were significantly different (P < 0.01).

Resistance to azithromycin and erythromycin of macrolides class

One study determined the susceptibility to azithromycin and erythromycin, including four isolates, all of which were found to be resistant, with an RR of 100% (95% CI 89%−100%). Only one data set from 2010 to 2015 in North America was available (Figs. 3 and 4).

Resistance to clindamycin of lincosamides class

One study determined the susceptibility to clindamycin, including eight isolates, all of which were found to be resistant, with an RR of 99% (95% CI 92%−100%). Only one data set from 2010 to 2020 in Oceania was available (Figs. 3 and 4).

Resistance to peptide antibiotics class

Bacitracin

One study determined the susceptibility to bacitracin, including eight isolates, all of which were found to be resistant, with an RR of 99% (95% CI 92%−100%). Only one data set from 2010 to 2020 in Oceania was available (Figs. 3 and 4).

Polymyxin B

One study determined the susceptibility to polymyxin B, including eight isolates, seven of which were found to be resistant, with an RR of 86% (95% CI 62%−100%). Only one data set from 2010 to 2020 in Oceania was available (Figs. 3 and 4).

Discussion

P. aeruginosa is a prevalent opportunistic pathogen found in various environments and on biological surfaces [1, 2]. In equine veterinary medicine, infections with P. aeruginosa are relatively rare [4] but pose significant treatment challenges and poor prognosis once contracted. The pathogen’s intrinsic resistance to a broad spectrum of antimicrobials, combined with its capacity to acquire new resistances, leads to a complex array of resistance mechanisms and widespread AMR [11]. P. aeruginosa is well known for its capacity to form biofilms, a complex aggregation of microorganisms embedded within a self-produced extracellular matrix [35–37]. Biofilm formation significantly enhances bacterial resistance to antimicrobials and host immune responses [38–41]. Within biofilms, the diffusion of antimicrobials is inhibited, metabolic activity is reduced, and cells exhibit a persistent phenotype, all of which collectively diminish the efficacy of antimicrobial therapy [42–45]. The mechanisms of biofilm play a crucial role in chronic and recurrent equine infections, presented in wounds, cornea, reproductive tract and other tissues [20, 25, 26, 30, 46–48]. Despite its importance, biofilm-associated resistance remains under-investigated in equine veterinary medicine and warrants further research. The restricted range of antimicrobials approved for veterinary use can exacerbate resistance development over extended application [1]. Horses exhibited high sensitivity to many antimicrobials, which further limits the number of drugs authorized for clinical use. Therefore, the off-label antimicrobial use is relatively common in equine clinical practice, which not only increases the risk of adverse drug reactions, but also raises the risk of antimicrobial misuse and the subsequent development of resistance. Donkeys, as another key domesticated equine species, may also encounter similar issues. A study reports that 10.6% of bacteria isolated from endometrial infections in donkeys are P. aeruginosa [7]. Therefore, it is crucial to identify the precise AMR patterns of P. aeruginosa in equine sources.

In this study, we examined the AMR patterns of 1624 strains of P. aeruginosa isolated from 10 articles to clarify the resistance profile of this bacterium. In human medicine, a variety of antimicrobials are available for treating P. aeruginosa, each with different mechanisms of action. Although these antimicrobials have shown good therapeutic effects, resistance has emerged [49]. In the field of veterinary medicine, the emergence of AMR appears to be more prevalent and severe, as our analysis from this article indicated. All included studies conducted in vitro antimicrobial susceptibility tests, encompassing a range of drugs that are not commonly used or are not legally permitted for use in equine animals. Despite this, statistical findings indicate that P. aeruginosa isolates from equine sources still exhibit significant resistance to many non-first-line drugs and even drugs that are not allowed for use. This could be attributed to P. aeruginosa being an environmental bacterium that may interact with bacteria causing human infections. Additionally, it may be due to the limited number of drugs approved for use in equine animals, leading to potential non-compliant medication practices in clinical settings.

Results indicated that P. aeruginosa isolates were almost completely resistant to penicillins class antimicrobials such as oxacillin and amoxicillin, carbapenems class ertapenem, and cephalosporins class including cefazolin, cefoxitin, cephalexin, and cefadroxil. Furthermore, extensive resistance was observed against aminoglycosides class kanamycin, tetracyclines class tetracycline, macrolides class azithromycin and erythromycin, lincosamides class clindamycin, and peptide antibiotics class bacitracin. In contrast, the most effective antimicrobials were found to be monobactams class aztreonam, cephalosporins class ceftazidime, cefepime, and ceftolozane-tazobactam, phosphonic acids class fosfomycin, and fluoroquinolones class ciprofloxacin.

When considering antimicrobial classes based on our research findings, macrolides (RR 99%), lincosamides (RR 99%) and tetracycline (RR 99%) exhibit the highest resistance rates, followed by rifamycins (RR 93%), peptide antibiotics (RR 93%), and sulfonamides (RR 89%), potentially posing challenges for the treatment of infected animals. In contrast, monobactams (RR 1%), phosphonic acids (RR 1%), and fluoroquinolones (RR 23%) have the lowest resistance rates among the classes. Studies in human medicine have shown that the minimum inhibitory concentration of fluoroquinolones against P. aeruginosa increases with the escalating use of these antimicrobials, ultimately leading to the development of resistance [50, 51]. However, in the equine veterinary medicine, our findings indicated a decline in resistance rates for marbofloxacin of fluoroquinolones and doxycycline of tetracyclines in recent years, suggesting they could be a suitable treatment option. Nevertheless, their use should be approached with caution to prevent the accumulation of resistance.

A significant rise in resistance is noted for certain antimicrobials, such as penicillin, ampicillin, and amoxicillin-clavulanic-acid in the penicillins class (RR 78%−93%), cefquinome and ceftiofur in the cephalosporins class (RR 39%−52%), neomycin, netilmicin, and gentamicin in the aminoglycosides class (RR 29%−71%), enrofloxacin in the fluoroquinolones class (RR 32%), chloramphenicol in the amphenicols class (RR 82%), doxycycline and oxytetracycline in the tetracyclines class (RR 65%−87%), and polymyxin B in the peptide antibiotics class (RR 86%).

Additionally, the analysis reveals that resistance to imipenem, ceftiofur, and amikacin has significantly increased over time (from 2006 to 2016 to 2016–2023), likely due to the higher usage of these antimicrobials. This trend suggests a concerning reliance on last-resort human antimicrobials in equine medicine. As pathogens develop resistance to first-line veterinary antimicrobials, clinicians may be increasingly forced to resort to off-label use of carbapenems to manage recalcitrant infections. Although this approach may be clinically justified in individual cases, the risk of accelerating the emergence and spread of carbapenem-resistant P. aeruginosa at the population level. Conversely, resistance to gentamicin, marbofloxacin, doxycycline, chloramphenicol, and trimethoprim-sulfamethoxazole has significantly decreased, which may suggest a reduction in the usage of these antimicrobials.

Geographic analysis reveals significant regional variation in AMR prevalence. Asia demonstrates the highest resistance rates (average RR 80%), followed by North America (74%) and Oceania (70%). Europe shows the most favorable profile with substantially lower resistance rates (average RR 25%). This distinct regional disparity is likely driven by divergent regulatory frameworks and antimicrobial stewardship practices. In Europe, strict regulations enforce the prudent use of antimicrobials and restrict the veterinary use of critical human drugs. Conversely, in parts of Asia, the higher resistance rates may reflect less stringent enforcement of prescription-only policies and easier access to broad-spectrum antimicrobials without veterinary oversight. This regulatory gap creates a stronger selective pressure in these regions, accelerating the dissemination of resistant clones. Additionally, there is a complete lack of resistance data from South America, which warrants further exploration. Among β-lactam combination agents, amoxicillin-clavulanic acid has the highest resistance rate, indicating the need for monitoring its usage. However, the current evidence is largely limited to Asian populations, highlighting the necessity for expanded surveillance across other geographic regions. Imipenem, a carbapenem, has a relatively low overall resistance rate (RR 8%), but its resistance has surged in recent years, particularly in North America, indicating a need for restricted use of this antimicrobial in that region. Amikacin, an aminoglycoside (RR 11%), is also of concern, with resistance predominantly in Asia (RR 87%), followed by North America (RR 25%). Similarly, ceftiofur in the cephalosporins class warrants attention due to its overall higher resistance rate (RR 52%), with resistance predominantly in Europe (RR 99%), followed by Asia (RR 69%) and Oceania (RR 88%). For trimethoprim-sulfamethoxazole, a decline in resistance in recent years is likely due to effective control measures in Asia (RR 70%), as resistance rates in other regions are extremely high (RR 98%−100%). Similarly, the decrease in resistance to chloramphenicol appears to be due to effective control in regions other than North America, where resistance is at 100%. These findings demonstrate the success of controlling resistance rate in these regions and underscore the need for intensified antimicrobial stewardship efforts in areas with persistently high resistance rates.

Mechanistically, the high resistance rates observed across various antimicrobial classes in P. aeruginosa can be attributed to a complex interplay of intrinsic, acquired, and adaptive resistance mechanisms [52, 53]. The production of β-lactamases by P. aeruginosa represents a significant component of its intrinsic AMR, conferring exceptionally high resistance to most β-lactam antimicrobials [11]. To counteract this resistance, β-lactamase inhibitors such as clavulanate and tazobactam have been developed and have been found to substantially enhance the efficacy of β-lactam antimicrobials [54]. This study revealed that while amoxicillin demonstrates high-level of resistance (RR 99%), its combination with clavulanic acid significantly enhances antimicrobial activity. Nevertheless, the amoxicillin-clavulanic acid combination still exhibits substantial resistance (RR 78%). One notable finding is the marginally higher resistance rate to ticarcillin–clavulanic acid (RR 8%) compared to ticarcillin alone (RR 7%). Although seemingly paradoxical given clavulanic acid’s role as a β-lactamase inhibitor, this phenomenon can be explained by the frequent production of clavulanic acid-resistant β-lactamases in P. aeruginosa, particularly AmpC-type and metallo-β-lactamases. In these cases, the inhibitor may not only fail to enhance therapeutic efficacy but may also potentially exert selective pressure favoring the proliferation resistant strains [55–57]. Moreover, the small sample size in this subgroup may exaggerate minor differences in resistance rate. Further investigation with larger datasets and specific β-lactamase profiling is warranted to clarify this phenomenon.

Carbapenems were initially developed to combat Gram-negative bacteria that produce β-lactamase and are resistant to broad-spectrum penicillin [49]. However, P. aeruginosa exhibits intrinsic resistance to carbapenem antimicrobials through other mechanisms. Notably, carbapenems can be categorized into two groups based on their clinical spectrum and pharmacological properties. Group 1 carbapenems, such as ertapenem, are primarily indicated for the treatment of infections caused by non-pseudomonal, non-Acinetobacter Gram-negative pathogens in humans, including those resistant to other antimicrobial classes. In contrast, Group 2 carbapenems—such as doripenem, imipenem, and meropenem—are more appropriate for managing infections where P. aeruginosa is involved due to their broader spectrum of activity [58, 59]. In our study of equine-associated P. aeruginosa, we observed a marked difference in resistance between these two carbapenems: ertapenem exhibited the highest resistance rate (RR 100%), while imipenem had the lowest (RR 8%). This disparity can be attributed to the intrinsic pharmacodynamic and structural differences between the drugs. Specifically, ertapenem demonstrates limited activity against P. aeruginosa because of its weak binding affinity for the bacterium’s penicillin-binding proteins (PBPs) and reduced ability to penetrate the outer membrane [60, 61]. In contrast, imipenem shows stronger PBP binding and better membrane permeability, contributing to its enhanced effectiveness [62–65]. Furthermore, differences in susceptibility to efflux pump activity and porin channel modifications may also contribute to the observed resistance pattern between these agents [66–70]. However, annual subgroup analysis indicates a rising trend in imipenem resistance, likely driven by increased clinical usage, highlighting the urgent need for strict antimicrobial stewardship.

Beyond carbapenems, P. aeruginosa also demonstrates diverse resistance mechanisms to other antimicrobial classes commonly used in veterinary practice. Aminoglycoside resistance is driven mainly by the production of aminoglycoside-modifying enzymes (AMEs), which enzymatically alter the drugs and prevent them from binding to the bacterial ribosomes, and by 16 S rRNA methylation or efflux-based mechanisms [71, 72]. Fluoroquinolone resistance is largely associated with mutations in the quinolone resistance-determining regions (QRDRs) of the gyrA and parC genes, encoding DNA gyrase and topoisomerase IV, respectively. Additionaly, efflux pumps such as MexEF-OprN contribute to resistance by reducing intracellular drug accumulation [73, 74]. Tetracycline resistance is mediated by both ribosomal protection proteins and efflux pumps encoded by tet genes, which block the antimicrobial’s ability to inhibit protein synthesis [75, 76]. For macrolides, lincosamides, and polymyxins, P. aeruginosa exhibits intrinsic resistance due to the low permeability of its outer membrane, the lack of compatible ribosomal binding sites, and the activation of resistance operons such as arnBCADTEF that modify lipid A and reduce drug affinity [77–80]. Additionally, resistance to trimethoprim-sulfamethoxazole is typically due to target site mutations or acquisition of plasmid-borne genes that encode resistant dihydrofolate reductase or dihydropteroate synthase [81–83]. Chloramphenicol resistance involves enzymatic inactivation by chloramphenicol acetyltransferases and active drug efflux [84–86]. Although fosfomycin resistance in P. aeruginosa is often mediated by fosA genes and reduced transporter activity, its relatively low resistance rate may reflect its limited veterinary use [87]. Together, these diverse resistance mechanisms underscore the adaptive capacity of P. aeruginosa and highlight the urgent need for molecular-level monitoring of AMR in equine pathogens.

This study acknowledges certain limitations. First and foremost, it is important to note that some drugs are only mentioned in a single study, which raises questions about their reliability. However, they still hold some reference value. The status of these drugs is explained in the main text, or they lack data for heterogeneity analysis in the data analysis tables, involving piperacillin, piperacillin-tazobactam, oxacillin, amoxicillin, ertapenem, aztreonam, ceftazidime, ceftolozane-tazobactam, cefazolin, cefoxitin, cephalexin, cefadroxil, fosfomycin, netilmicin, tobramycin, kanamycin, neomycin, levofloxacin, oxytetracycline, azithromycin, erythromycin, clindamycin, bacitracin and polymyxin B. These details can be used to distinguish them from other drugs. It is encouraged to approach these data with a critical mindset. Besides, the analysis is based on a limited number of articles and strains, particularly for donkeys, which may not be sufficient to fully represent the AMR profile of P. aeruginosa across all equine species. The study’s timeframe spans from 1990 to 2024, which is extensive, but data may be sparse in certain periods, potentially affecting the analysis of resistance trends. Additionally, there is a scarcity of data for specific antimicrobials in some regions, and varying conditions and standards may lead to high heterogeneity among the studies involved. Therefore, a more comprehensive understanding and analysis of the AMR of equine-associated P. aeruginosa will depend on future research and surveillance from various regions.

Conclusions

In conclusion, the stark reality of AMR among equine-associated P. aeruginosa is alarming and underscores a growing threat to equine healthcare globally. This systematic review and meta-analysis revealed high resistance rates to commonly used antimicrobial classes, including penicillins, macrolides, tetracyclines, and aminoglycosides, with resistance rates in some cases exceeding 90%. In contrast, lower resistance was observed for select agents such as aztreonam, ceftazidime, fosfomycin, and ciprofloxacin, although these agents showed regional variability. Geographically, the problem appears most concerning in Asia, where resistance rates are highest across multiple drug classes, likely reflecting unregulated antimicrobial use. Conversely, lower resistance in Europe may indicate more successful implementation of stewardship and regulatory interventions. These findings highlight the critical need for targeted antimicrobial stewardship in veterinary practice, region-specific resistance surveillance, and further research into novel therapeutic options. Without coordinated global efforts, the treatment of P. aeruginosa infections in equine may become increasingly limited. Strengthening antimicrobial policies and promoting judicious antimicrobial use are essential to preserve the efficacy of current therapies and safeguard equine and public health.

Authors’ contributions

All authors contributed substantially to this work. Jing Li and Yiping Zhu concepted the study. Luo Yang and Yuxin Xie conceived the study and drafted the manuscript. Guangzhi Zhong conducted the data analysis. Jing Li, Yiping Zhu and Dejun Liu reviewed the manuscript. All authors have read and agreed upon the current version of the manuscript.

Funding

This systematic review was funded by National Natural Science Foundation of China (Grant No. 32202861).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.