Introduction
Extended-spectrum β-lactamase-producing Enterobacterales (ESBL-PE) are a serious global health concern for transmission of multidrug-resistant organisms, particularly Escherichia coli and Klebsiella pneumoniae. Hospital-acquired infections, including surgical site infections caused by ESBL-PE, are associated with considerable morbidity and mortality [Reference Horan, Gaynes and Martone1]. Contaminated surgical wounds and medical devices, along with admission to hospital more than 24 hours before surgery, were identified as the most statistically significant risk factors in a recent study [Reference Sawyer and Evans2] and underline the need for preventive measures to improve surgical outcomes [Reference Banerjee and Johnson3].
We investigated whether faecal carriage of ESBL organisms in patients before abdominal surgery constituted a source of post-surgical infections in these subjects. Isolates recovered from rectal swabs of 104 patients 1 day before and up to 3 days post-surgery were characterized by molecular characteristics and ESBL resistance genes to confirm prior colonization with and persistence of ESBL-PE strains.
Materials and methods
Rectal swabs were cultured on selective CHROMagar ESBL (SIGMA-ALDRICH, St. Louis, USA) and MacConkey agar (Becton, Dickinson, Sparks, USA). Isolates were identified to the species level by standard biochemical tests. Combination disk diffusion tests [Reference Banerjee and Johnson3] were performed for phenotypic confirmation of the presence of ESBLs using appropriate control reference strains. ESBL production was confirmed by an increase of ≥5 mm using combination disks of cetazidime (30 μg)/clavulanate (10 μg) or cefotaxime (30 μg)/clavulanate (10 μg) compared against CAZ (30 μg), or CTX (30µg) alone. K. pneumoniae ATCC 700603 and E. coli ATCC 25922 were included as ESBL-positive and ESBL-negative controls.
Detection of resistance genes by whole-genome sequencing
Of the 104 patients who underwent abdominal surgery from July 2018 to March 2019, 31 were positive for ESBL-producing E. coli and K. pneumoniae in their faecal flora. From the 17 patients who yielded ESBL-PE organisms on pre- and post-surgical screening, 46 isolates were recovered, except for one patient (number 30) where ESBL-KP and KPR phenotypes were found only in the post-operation specimen. The 46 selected isolates were subjected to whole-genome sequence analysis.
The quantity and quality of DNA extracts were determined by gel electrophoresis and fluorescent measurement by Qubit assay (Thermo Fisher Scientific, Vilnius, Lithuania). DNA libraries were constructed using MGIEasy FS DNA library kit and sequenced with a DNBSEQ-G400 sequencer (MGI Tech, Shenzhen, China). All isolates underwent a quality control process. Reads with a mean quality score <Q30 or length <36 base pairs were discarded. KRAKEN2 (v2.1.2) [Reference Wood, Lu and Langmead4] was used to remove unclassified reads, and de novo assembly was performed with Unicycler (v0.5.0).
Multilocus sequence types (STs) and taxa were identified by MLST (v2.11) [Reference Jolley and Maiden5] and KRAKEN2, respectively. Drug-resistant genes were identified with two approaches. First, passed quality reads were mapped to the ResFinder database with ResFinder (v4.1.11). Second, assembled contigs were mapped to NCBI AMRFinderPlus and ResFinder databases. In this approach, ABRicate (v1.0.1) [Reference Seemann6] was used to map assembled sequences to the NCBI AMRFinderPlus database [Reference Feldgarden, Brover and Haft7] and ResFinder database [Reference Zankari, Hasman, Cosentino, Vestergaard, Rasmussen, Lund, Aarestrup and Larsen8]. All resistance genes identified were pooled, and those with more than 90% identity and coverage were selected. Additionally, plasmid replicons were identified using PlasmidFinder 2.0 (https://cge.cbs.dtu.dk/services/PlasmidFinder/).
The accession number of the isolates is PRJNA1020534 (release date: 2024-2103-01, https://www.ncbi.nlm.nih.gov/sra/PRJNA1020534); BioProject and associated SRA metadata are available at https://dataview.ncbi.nlm.nih.gov/object/PRJNA1020534?reviewer=7v8hhrtb4ago7ut598acfeq5gc.
Results
All isolates were identified to species level as E. coli and K. pneumoniae classified by the sequence taxonomic database. Species classifications were confirmed to be correct except for two isolates SK106 and SK128 from patient numbers 24 and 29, respectively, which were reassigned from E. coli to Enterobacter roggenkampii (EER) (Supplementary Table S1). All but 7 isolates were assigned to an ST, and in total, 23 different STs were identified (Supplementary Table S1). E. coli isolates exhibited the greatest heterogeneity with 20 STs, whereas K. pneumoniae isolates were mainly ST14 and ST15. Almost all isolates from pre- and post-surgical samples shared the same ST, and isolates from 8 of the 17 patients fell in the same type. Notably, the K. pneumoniae isolated from patient number 30 belonged to the same genotype (ST14) as with others of this patient’s isolates but harboured different resistance genes.
Sequence analysis revealed the presence of bla genes in addition to other resistance genes. The most prevalent bla ESBL genes in E. coli were bla EC (69.57%), bla CTX-M (65.22%), and bla TEM (36.95%), whereas bla SHV predominated in K. pneumoniae (41.30%) alone (Supplementary Table S2). The bla genes found from the isolates at pre- and post-surgery were generally of the same prevalence. Most patients, except for six individuals, had almost the same resistance gene profiles of isolates pre- and post-abdominal surgery (Supplementary Table S2).
Three classes of β-lactamases were identified: class A (bla CTX-M, bla TEM, bla Z), class C (bla SHV, bla MIR, bla DHA), and class D (bla OXA). The most prevalent bla CTX-M, bla TEM, bla SHV, and bla OXA variants were bla CTX-M-15, bla TEM-1B, bla SHV-28, and bla OXA-1, respectively. All K. pneumoniae strains harboured bla SHV with seven different variants, namely bla SHV-11, bla SHV-13, bla SHV-28, bla SHV-100, bla SHV-106, bla SHV-110, and bla SHV-187 (Table 1).
Distribution of bla genes among 46 strains isolated from rectal swab of patients at pre- and post-abdominal surgery
Interestingly, pAmpC β-lactamase genes, including bla CMY-2, were found in both E. coli and K. pneumoniae and bla DHA-1 in the latter. Moreover, mcr-10.1, mobile colistin resistance genes, were detected only in resistant Enterobacter cloacae (ECLR) (SK131 and SK132). The ESBL-producing EER carried bla TEM-1B, bla CTX-M-3, and bla MIR-2.
Various plasmid types harbouring antimicrobial resistance genes were identified such as IncFIA, IncFIB, IncFIC, IncFII, IncQ1, IncX1, IncHI2, and IncR (accession number: PRJNA1020534).
Discussion
Different database or input sequence formats were used for sequence data analysis, leading to different drug resistance identification results. Consequently, multiple databases were used to provide more accurate data. For example, ResFinder identified the SHV gene when using non-assembled sequences as input, but this gene was not flagged when using assembled sequences in samples SK116, SK125, SK126, and SK127. It is possible that the process was unable to assemble the SHV sequence due to the known performance limitation of de novo assembly on short read data. However, the unknown ST and missing taxonomy classifications were recalled from KRAKEN2, which contained multiple taxonomical profiles of various species. Long read sequence data is therefore necessary for further approaches.
bla CTX-M, bla TEM and bla SHV are the most prevalent of the many ESBLs detected in various pathogens, and consequently, they have become widely disseminated across various epidemiological niches. A previous study found SHV to be distributed mostly among K. pneumoniae [Reference Ur Rahman, Ali and Ali9], and here, it was found only in this species. However, variants of the SHV type have been detected in other members of the Enterobacterales family and Acinetobacter baumannii [Reference Naas, Namdari and Réglier-Poupet10, Reference Liakopoulos, Mevius and Ceccarelli11].
In this study, the presence of the same ST types of strains present at pre- and post-surgery was interpreted as being indicative of colonization of the patient’s gut by ESBL producers and other resistant strains before surgery. Plasmid-mediated resistance genes are readily transferable and often spread from one bacterium to another. It follows that the persistence of such strains can give rise to hazardous and difficult-to-treat post-surgical site infections. Hence, screening of patients before, and after, surgery to confirm persistent carriage of ESBL-PE strains is of practical benefit and increases clinical awareness of their potential transmission during surgery.
The multi-resistant EPEC ST131 strain has been reported worldwide due to its high risk of gastrointestinal tract infection and sometimes progression to urinary tract infection and septicaemia. It is also widely distributed as a colonist among healthy individuals and animals [Reference Banerjee and Johnson3, Reference Pitout and DeVinney12, Reference Doi, Iovleva and Bonomo13]. This genotype is particularly associated with several resistance genes, particularly bla CTX-M [Reference Doi, Iovleva and Bonomo13]. The isolates harbouring bla CMY-2, which is the most common pAmpC β-lactamase gene reported worldwide [Reference Jacoby14], and mcr-10.1 present a potentially high risk of infections during abdominal surgery in this study.
Colistin was only relatively recently introduced as the last available antibiotic for combatting multiple drug-resistant bacterial infections [15], but the presence of its resistance gene, mcr, in this study indicates that genetically mobilized colistin-resistant strains pose an emerging threat due to their associated high risk of morbidity and mortality. Variants of the mcr gene including mcr-1 through mcr-10 have been identified in many bacteria globally [Reference Li, Nation and Turnidge16].
In patient number 24, an EPEC strain was isolated before surgery and an ESBL-producing EER after surgery. Both isolates were positive for bla TEM-1B and bla CTX-M-3. These genes and bla MIR-1, a plasmid-mediated class C (group 1), confer resistance to oxyimino β-lactams. They were detected in EER, while bla CMY-2 was found in EPEC. The presence of the plasmid-mediated genes of the two species may result in their potential transfer between the strains during intestinal carriage. It is widely accepted that appropriate antibiotic use for prophylaxis is essential to reduce infections in high-risk patients. Likewise, guidelines for appropriate drug prescriptions for such individuals should be evaluated, and patients should be actively screened for carriage of ESBL producers and other resistance genes before surgery.
Extended-spectrum β-lactamase producers were not detected in 120 healthy adults as previously reported from a tertiary Thai hospital [Reference Pongpech, Naenna and Taipobsakul17]. However, ESBL-producing E. coli and K. pneumoniae multidrug-resistant isolates were recently reported in approximately 30% of an elderly population living at home who had undergone abdominal surgery [Reference Aksarakorn Kummasook, Nuanmuang and Baiubon18].
In conclusion, phenotypic and genotypic characteristics of a collection of isolates of ESBL-producing E. coli and K. pneumoniae and other plasmid-mediated resistant strains, especially mobilized colistin resistance gene mcr, is necessary to arrest their potential spread. This study provided detailed information on the species distribution and their resistance genes, which will aid prevention and control of post-abdominal surgical infections, and the spread of resistance genes.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0950268824000578.
Data availability statement
The authors confirm that the data supporting the findings of this study are available within the article.
Acknowledgements
Our great appreciation goes to the surgical team at Thammasat University Hospital (Dr. C. Mingmalairak, Dr. P. Mahawongkajit, Dr. J. Juntong, Dr. P. Limpavitayaporn, Dr. E. Sriussadaporn, Dr. A. Tongyoo, Dr. P. Boonyasatid, Dr. T. Chunsirisub, Dr. K. Nakornchai, and Dr. W. Thowprasert) for facilitating specimen collection. I am grateful to Dr. S. Trakulsomboon and Dr. Pattarachai Kiratisin for providing positive control strains. I would like to extend my sincere thanks to the Faculty of Medicine, Thammasat University and Information Technology Unit and Thammasat Research & Innovation Unit, Thammasat University Hospital.
Author contribution
S.K. designed and directed the research project, and S.K. and L.N. performed experiments under the supervision of S.K. S.K., P.Ph., and A.A. P.Pa. contributed to important comments. W.P., L.N., W.R., and P.Ph. carried out bioinformatics analyses, and S.K. wrote the main manuscript with review input from all authors.
Funding statement
This research was supported by the Thammasat University Fund, Contract No. TUFT 6/2566 and General Research Fund, Faculty of Medicine, Contract No. 2–22/2566. The poster presentation on this research was funded by the Thammasat University Travel Fund.
Ethical standard
This study was approved by the Human Research Ethics Committee No. 1, Faculty of Medicine, Thammasat University, current name: Human Ethics Committee of Thammasat University (Medicine), Thailand (MTU-EC-DS-2-014/61). All patients consented to the study.
Open access
