Resistance of Enterobacteriaceae to extended-spectrum-beta-lactamase (ESBL) has spread rapidly worldwide and poses a serious threat in many regions [1]. Klebsiella pneumoniae (K. pneumoniae) is one of the most common ESBL-producing Enterobacteriaceae in healthcare-associated infections (HAI) including hospital infection (HA) and community infection (CA). In recent years, the prevalence of ESBL-producing K. pneumoniae in clinical settings has been accelerated by the inappropriate application of antibiotics leading to a significant public health concern [2, 3]. ESBLs are transmitted by a resistance plasmid or integron in, and between, hospitals among Enterobacteriaceae, frequently resulting in multiple antimicrobial resistance and clinical treatment failures. The nosocomial infections of K. pneumoniae frequently occur due to horizontal transfer of resistant mobile genetic elements in hospitals and in the community, which contributes to multiple mechanisms of transmission of antimicrobial agents [4, 5]. Therefore, understanding the molecular mechanism of resistance to antibiotics and predicting the clinical characteristics of antimicrobial drugs, we investigated the occurrence and rapid dissemination of K. pneumoniae producing ESBLs from obstetrics and gynaecology settings in three associated hospitals of a university over a three-year period in China to make a better choice of antimicrobial agents and improve patient survival.

This retrospective study was based on a survey of 6147 in-patients under obstetrics & gynaecology settings in three associated hospitals of a university over a three-year period in China. Patients with concurrent HAIs were classified into respiratory tract infections including pneumonias, bloodstream infections including catheter-related infections; urinary tract infections, surgical site infections, and other infections were excluded. For purposes of data analysis, we collected data for each HAI including age, date of onset, and microbes. The pregnant women admitted to obstetrics settings had an average age of 36.7 $\pm$ 3.5 years with high maternal age for caesarean section operations. The women admitted to gynaecology settings had a median age of 63 years for treatment or surgery of a cancer or tumour. Only gram-negative aerobic and facultative anaerobic isolates from infection sites were considered clinically significant. The pathogenic bacteria associated with HAI were collected within 48 h or after 48 h of hospitalisation according to local hospital criteria. The research was approved by the institutional ethics committees of the participating hospitals and written informed consent was obtained from all participants enrolled in the study.

A single specimen was obtained from sputum or bronchial washing. All isolates were identified as K. pneumoniae by the conventional standard procedures and confirmed by VITEK GNI system (bioMérieuk Vitek Inc., USA). No repetitive isolates from a single patient were included. Susceptibility to 16 antimicrobial agents was determined and interpreted by an agar doubling dilution method akin to that in the Clinical and Laboratory Standards Institute (CLSI) criteria. ESBL screening was determined by using discs (Bio-Rad, Hercules, CA, USA) according to CLSI guidelines [6]. Control strains were E. coli ATCC 25922, Streptococcus pneumoniae ATCC 49619, and P. aeruginosa ATCC 27853.

A series of bla genes including bla${}_{T\mathit{}E\mathit{}M}$, bla${}_{S\mathit{}H\mathit{}V}$, bla${}_{O\mathit{}X\mathit{}A}$, and bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M}$ were identified by PCR amplification, primers and PCR reactions were performed as described elsewhere [7, 8]. PCR products of bla genes were sent to Sangon Biotech Co., Ltd (Shanghai, China) for sequencing, and DNAman software was used to analyse the sequencing results.

From January 2017 to December 2019, we calculated that a total of 1564 patients would be recorded to estimate an anticipated HAI prevalence of 21.8%. A total of 1564 isolates, which were obtained from respiratory tract and sputum, frequently included Klebsiella species (33.1%), P. aeruginosa (27.8%), Acinetobacter species (17.8%), Enterococcus species (10.9%), and Escherichia coli (9.3%). Besides that, others were gram-positive including Staphylococcus species (9.8%, including 3.7% S. aureus).

In total, out of 518 K. pneumoniae isolates, 124 strains were isolated from pregnant patients with an average hospitalisation length of 10 days for caesarean section operations, and 394 strains were separated from patients for treatment or surgery of a cancer or tumour. The average duration of hospitalisation exceeded 15 days. 358 isolates (69.2%) produced ESBLs, and 160 isolates (30.8%) did not produce ESBLs based on antimicrobial susceptibility testing. In view of the distribution of the three hospitals, 222 K. pneumoniae isolates were identified in the first affiliated hospital. 158 (158/222, 71.2%) were ESBL producers and 64 (64/222, 28.8%) were not; 114 isolates (114/174, 65.5%) were ESBL producers and 60 strains (60/174, 34.5%) were not in 174 isolates of K. pneumoniae from the second affiliated hospital; additionally, out of 122 isolates of K. pneumoniae, 86 (86/122, 70.5%) were ESBL producers and 36 (36/122, 29.5%) were not in the third affiliated hospital.

All ESBL-producing Klebsiella pneumoniae isolates were resistant to aztreonam and cefpodoxime, and the least resistant antibiotic was ceftazidime (67.8%). Besides gentamycin, as a non-beta-lactam, imipenem was also found to confer effectiveness as an antibiotic. In non-ESBL K. pneumoniae strains, resistance to aztreonam was not observed, however, the highest resistance to cefotaxime and ceftazidime were observed (more than 50%) (Table 1).

All ESBL-producing isolates were subjected to PCR for detection of the most widely spread ESBLs: genes and amplicons were subjected to sequencing. Among K. pneumoniae-producing ESBLs, 88 strains (88/358, 24.6%) contained the bla${}_{T\mathit{}E\mathit{}M\mathrm{-}\mathrm{2}}$ gene, and 74 strains (74/358, 20.1%) were positive for the bla${}_{S\mathit{}H\mathit{}V\mathrm{-}\mathrm{3}}$ gene, 44 strains (44/358, 12.3%) contained the bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{2}}$ gene, 42 strains (42/358, 11.7%) contained the bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{4}}$ gene, 38 strains (38/358, 10.6%) contained the bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{5}}$ and bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{8}}$ genes, respectively, 34 (34/358, 9.5%) strains contained the bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{15}}$ gene. Also, 20 (20/358, 5.6%) strains contained bla${}_{O\mathit{}X\mathit{}A\mathrm{-}\mathrm{1}}$. The co-existence of bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M}$ and bla${}_{S\mathit{}H\mathit{}V\mathrm{-}\mathrm{2}}$ was identified in 20 strains (20.9%), 12 strains (13.4%) co-existed with bla${}_{S\mathit{}H\mathit{}V\mathrm{-}\mathrm{3}}$ and bla${}_{T\mathit{}E\mathit{}M\mathrm{-}\mathrm{2}}$, 14 isolates (6%) harboured both bla${}_{T\mathit{}E\mathit{}M\mathrm{-}\mathrm{2}}$ and bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{15}}$, eight isolates (6%) contained bla${}_{S\mathit{}H\mathit{}V\mathrm{-}\mathrm{3}}$, bla${}_{T\mathit{}E\mathit{}M\mathrm{-}\mathrm{2}}$, and bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{15}}$ (Table 2).

ESBL-producing K. pneumoniae are common pathogens isolated from various body sites. Our findings showed that the frequency of HAIs and antimicrobial resistance in K. pneumoniae isolates posed a considerable threat to public health in the region. In these obstetrics and gynaecology settings, pregnant patients for caesarean section operations, female patients treated for cancer or tumour were more susceptible to infection than those in other departments due to the length of hospitalisation, age, use of retention tube, the low patient-to-nurse ratio, and ICU bed occupancy. The findings may be explained by the fact that these patients were administered an excess of antibiotics due to their immunological status. As shown elsewhere, the main sites of infection are the respiratory tract, incision, and urinary tract: these are the most likely sites to cause infection in gynaecological in-patients [9].

Among in vitro ESBL-producing susceptibility to the third generation of cephalosporins, K. pneumoniae isolates included in the study exhibited the highest resistance to aztreonam, cefpodoxime, aztreonam, ceftriaxone, as well as ceftazidime: however, all isolates were susceptible to imipenem, meropenem, and ciprofloxacin (92.3%). So, imipenem or meropenem prophylaxis can be administrated as an effective antibiotic for K. pneumoniae-producing ESBL isolates. Evidence indicated that the number of administrations of third-generation cephalosporins (except imipenem or meropenem) affected incidence of HAIs. This finding may be supported by the fact that these infecting isolates were resistant to the third-generation cephalosporins, while all of them were susceptible to imipenem or meropenem in addition to gentamycin. These findings correlate with those reported previously by authors in China where 59.2% of isolates were ESBL-positive and all isolates were susceptible to imipenem or meropenem [10].

In the present study, we found that the rates of ESBL production were variable among three hospitals. The data ranged from 71.2% in the first hospital to 65.5% in the third hospital, showing the percentage of K. pneumoniae isolates from respiratory infection that contributed to different frequencies of ESBL production. The highest rates of K. pneumoniae isolates producing ESBLs were exhibited in the first hospital, and the lowest in the third hospital. Our finding showed that the frequency of occurrence of bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M}$ exceeded that of bla${}_{T\mathit{}E\mathit{}M}$ and bla${}_{S\mathit{}H\mathit{}V}$. The prevalence of bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M}$, bla${}_{T\mathit{}E\mathit{}M}$, and bla${}_{S\mathit{}H\mathit{}V}$ genes in this study was 54.7%, 23.9%, and 16%, respectively. Our findings are in agreement with the results from a survey in which Guo et al. found that 31.3 % of K. pneumonia isolated from one of the largest hospitals in central China were ESBL-positive whereas their detected prevalence of bla${}_{T\mathit{}E\mathit{}M}$, bla${}_{S\mathit{}H\mathit{}V}$, bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}I}$, and bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}I\mathit{}I\mathit{}I}$ among these isolates was 97%, 71.6%, 73.1%, and 20.9%, respectively [11]. Our results showed high prevalence of bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M}$, bla${}_{S\mathit{}H\mathit{}V}$, and bla${}_{T\mathit{}E\mathit{}M}$ and a low frequency of occurrence of bla${}_{O\mathit{}X\mathit{}A\mathrm{-}\mathrm{1}}$. Significantly, bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M}$ was found to be the dominant ESBL in K. pneumoniae isolates. They exhibited a diverse genotype consisting of bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{2}}$, bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{4}}$, bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{5}}$, bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{8}}$, and bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{15}}$. The increasing proportion of bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M}$ in the different settings suggested that the dissemination of K. pneumoniae isolates carried ESBL genes from remote wards to local hospital. Additionally, bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{15}}$ was first reported in India in 1999: this ESBL-type gene has now propagated in various countries and has become globally predominant. Studies demonstrated that the bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M}$ alleles have evolved from bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{2}}$ to bla${}_{C\mathit{}T\mathit{}X\mathrm{-}M\mathrm{-}\mathrm{15}}$ through point mutation or substitution owing to overuse of cefotaxime and ceftriaxone, resulting in a temporal shift in ESBL type. Similar findings were reported by other authors in Italy, Iran, and Turkey [4, 12, 13]. Another finding was that K. pneumoniae-positive for all three genes were found. This may be due to the coexistence of three genes in the same plasmid or mobilisation of genetic elements.

The emergence and spreading dissemination of ESBL-producing strains is of concern to public health officials: the detection rate of ESBL-producing K. pneumoniae increased significantly in recent years. One of the reasons for this may be related to inadequate and excessive use of antimicrobial drugs, a lack of adequate antimicrobial surveillance programmes, and such in-patients are in poor health: if nursing care was not timely or indeed absent, they are easily infected. Moreover, incomplete sterilisation of the ward can cause bacterial accumulation, and patients suffer from respiratory tract infection. So, strengthening the training of obstetrics and gynaecology nursing staff have been deemed essential to improve their awareness of infection prevention and control [14].

Hospital management needs to be given enough attention to formulate effective measures to reduce the incidence of HAIs. The management system of infection should be completed, and each operational process should be standardised; strict aseptic operations should be performed, and the disinfection of various instruments and equipment should be done well during hand surgery. The ward was disinfected regularly to ensure that the patient was in a clean, healthy, and sterile environment [15, 16].

There are some limitations to the current study: we only speculated upon the K. pneumoniae isolates and ESBL genetic contents from an obstetrics and gynaecology setting. We lacked evidence from a hospital surveillance study aimed at detecting K. pneumoniae isolates. Additional research is needed to explore the specific enzymes and kinetics present in K. pneumoniae isolates producing ESBL in the region.

In summary, the high prevalence of ESBLs in K. pneumoniae isolates with diverse enzyme gene types indicated the need for screening the changes in ESBL-producing isolates in this region. The data in this study contributed to the increasing recognition of ESBL isolates, enhanced surveillance, and strict control of hospital epidemics in this species.

ZXM and LMC designed the present study. LMC, BXY and CJY performed the assay, analyzed and interpreted the data. WH wrote the draft manuscript. BXY and CJY collected clinical samples. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.

The research was approved by the institutional ethics committees of the participating hospitals (Protocol Number 2017-02-01), and written informed consent was obtained from all participants enrolled in the study.

Thanks to all the peer reviewers for their opinions and suggestions.

The present project was a part of National Natural Science Foundation of China project (81402979; 81541141), the Jilin and Zhejiang Science and Technology Development Program (20140307008YY and 2014C33155); Health and Family Planning Commission of Jilin Province (2018J098).

The authors declare no conflict of interest.