Clinical and molecular characteristics of Klebsiella pneumoniae infection in a tertiary general hospital of Wuhan, China

Objectives The aim of this study was to investigate the clinical and molecular characteristics of Klebsiella pneumoniae infection from a tertiary general hospital in Wuhan, China. Methods From December 2019 to August 2022, 311 non-duplicate isolates of K. pneumoniae were collected from a tertiary hospital in Wuhan. These comprised 140 carbapenem-resistant K. pneumoniae (CRKP) isolates and 171 carbapenem-susceptible K. pneumoniae (CSKP) isolates. The clinical characteristics of patients with K. pneumoniae infection were retrospectively collected. Polymerase chain reaction (PCR) assays were used to identify the main carbapenem resistance genes, virulence genes and multi-locus sequence typing (MLST) profiles of the isolates, and the Galleria mellonella infection model was used to determine their virulence phenotypes. Results Independent risk factors for CRKP infection were hypertension, neurological disorders, being admitted to the intensive care unit (ICU) and prior use of antibiotics. Patient with CRKP infection had higher mortality than those with CSKP infection (23.6% vs 14.0%, P < 0.05). One hundred and two sequence types (STs) were identified among the K. pneumoniae isolates, and the most prevalent ST type was ST11 (112/311, 36.0%). All of the ST11 isolates were CRKP. Among the 112 ST11 isolates, 105 (93.8%) harboured the carbapenem resistance gene blaKPC-2 (ST11-KPC-2), and of these isolates, 78 (74.3%, 78/105) contained all of the four virulence genes, namely rmpA, rmpA2, iroN and iucA, suggesting that these genes were widespread among the isolates responsible for K. pneumoniae infections. Conclusion In this study, ST11-KPC-2 was responsible for most of the K. pneumoniae infection cases. Carbapenem resistance rather than the co-occurrence of the virulence genes rmpA, rmpA2, iroN and iucA was associated with K. pneumoniae infection-related mortality during hospitalisation. Furthermore, a high proportion of ST11-KPC-2 isolates carried all of the four virulence genes.


Introduction
Klebsiella pneumoniae is a gram-negative, nonmotile, encapsulated opportunistic pathogen that generally colonises human mucous membranes, including the nasopharynx and gastrointestinal tract and is ubiquitously found in soils and the healthcare environment such as on the surfaces of medical devices [1,2].In immunocompromised patients, infants and older adults, it can cause a variety of infectious diseases, including lower respiratory tract infections, urinary tract infections, bacteraemia and sepsis, and can even lead to infection-related death [3][4][5].Furthermore, infection caused by drug-resistant K. pneumoniae, especially by carbapenem-resistant K. pneumoniae (CRKP), are known to be associated with longer hospital stays, higher medical costs and a higher risk of mortalities than those caused by carbapenem-susceptible K. pneumoniae [6].
Carbapenem is regarded as an effective antimicrobial agent for treating serious K. pneumoniae infections [7].However, carbapenem resistance has been found in K. pneumoniae isolates in many regions worldwide and has been recognised as a major public health concern due to its rapid transmission [8].In China, the average resistance rates of K. pneumoniae to imipenem and meropenem were 22.6% and 24.2%, respectively, in 2022, which were significantly higher than the rates of 2.9% and 3.0% reported in 2005 (http:// www.chine ts.com) [9].K. pneumoniae carbapenemase (KPC) and New Delhi metallo-β-lactamase (NDM) encoded by bla KPC and bla NDM , respectively, are the commonest factors responsible for carbapenem resistance in K. pneumoniae in most countries worldwide, whereas oxacillinase-48 (OXA-48) is most common in Africa [10][11][12].Moreover, co-occurrence of bla KPC-2 and bla NDM in K. pneumoniae has been reported in the USA, China and Brazil [13,14].A promising new β-lactamase inhibitor combination for treating infections caused by KPC-and NDM-co-producing K. pneumoniae is aztreonam-avibactam [15,16].
Hypervirulent CRKP (hv-CRKP) is a growing public concern due to its high drug resistance and virulence, posing a serious threat to human health [17].In 2018, in a hospital outbreak of ST11 hv-CRKP in Zhejiang province, China, the fatality rate was 100%.This was found to be attributable to the presence of pLVPK, a virulence plasmid (~ 170 kb long) containing two capsular polysaccharide regulatory genes (rmpA and rmpA2) and multiple siderophore gene clusters (iucABCD/iutA/iroBCDN), which imparted concurrent hypervirulence, multidrug resistance and high transmissibility.It was also shown that the occurrence of the virulence genes rmpA, rmpA2, iroN and iucA was correlated with the presence of the virulence plasmid pLVPK [18][19][20].Furthermore, an increasing occurrence of hv-CRKP is also being observed in other sequence types (STs), including ST15 and ST147 [21][22][23][24].Therefore, this study aimed to find the risk factors for and mortality associated with CRKP infection by comparing the clinical characteristics of CRKP infection cases with those of CSKP infection cases and to identify the molecular characteristics, namely, the presence of carbapenem resistance genes (bla KPC-2 , bla NDM like and bla OXA-48 like) and four virulence genes (rmpA, rmpA2, iroN and iucA), that are correlated with the resistance and virulence of CRKP isolates.
Patients with K. pneumoniae infection were included in the study and classified as follows: K. pneumoniae infection cases were classified as bacteraemia (n = 27, 8.7%) when a blood culture was positive for a K. pneumoniae strain and the patient had clinical signs of systemic inflammatory response syndrome; K. pneumoniae infections were classified as nonbacteraemia through nonblood cultures, typically for cases such as pneumonia (n = 151, 48.5%), urinary tract infection (n = 81, 26%), liver abscess (n = 9, 2.8%), abdominal infection (n = 16, 5.1%), cholecystitis (n = 8, 2.6%) and others (n = 19, 6.1%).Electronic medical records of the patients were retrospectively collected, and the following data were extracted: gender; age; smoking status; comorbidities such as diabetes, hypertension, liver abscess, respiratory failure, cancer (including haematological malignancies) and neurological disorders, use of antibiotics in the 3 months before the current hospitalisation and during the current hospitalisation until a positive K. pneumoniae culture was obtained; initial severity (in the ICU); immunosuppressive conditions containing immunosuppressives therapy, the carry of human immunodeficiency virus (HIV) and high dosage administration of corticoids; the coexistence of carbapenem-resistant Acinetobacter baumannii in the clinical specimen and mortality outcome of this hospitalisation.This study was approved by the Ethics Committee of Zhongnan Hospital, Wuhan University.

Detection of antimicrobial resistance and virulence genes
Genomic DNA was extracted from the K. pneumoniae isolates using a DNA extraction kit (Aidlab, Beijing) according to the manufacturer's instructions.All primers used in the present study are shown in Table 1.The presence of three carbapenemase genes, bla KPC-2 , bla NDM like and bla OXA-48 like, in the CRKP isolates was tested using a polymerase chain reaction (PCR) assay.The final volume of the PCR reaction mixture was 25 µL, which included 12.5 µL of 2 × buffer (Aidlab, Beijing), 9.5 µL ddH 2 O, 1 µL forward primer (10 µM), 1 µL reverse primer (10 µM) and 1 µL genomic DNA.The PCR amplification conditions were as follows: denaturation at 95 °C for 3 min, followed by 42 cycles of 95 °C for 30 s, 57 °C for 20 s and 72 °C for 20 s and final extension at 72 °C for 5 min.The PCR products were then subjected to 1.5-2.0%agarose gel electrophoresis for detection.The primers of the virulence genes rmpA, rmpA2, iroN and iucA were those used in a previous study [20].The annealing temperature and elongation time were determined according to the primer and product length.

Multi-locus sequence typing (MLST)
Seven house-keeping genes (infB, pgi, mdh, phoE, gapA, tonB and rpoB) were amplified from the genomic DNA of the K. pneumoniae isolates for MLST as described previously [25].The amplified products were purified and subjected to Sanger's dideoxy DNA sequencing (Tianyi Huiyuan, China).The resulting sequences were compared with the existing K. pneumoniae gene sequences available on the Pasteur Institute MLST website (http:// bigsdb.paste ur.fr/ klebs iella/).Each clinical isolate was then classified by the ST.

A Galleria mellonella model of K. pneumoniae infection
To test the virulence of the K. pneumoniae isolates, G. mellonella was purchased from Aidlab (China) and stored in the dark at 4 °C until testing.Overnight cultures of K. pneumoniae strains were washed with phosphate-buffered saline (PBS) and further adjusted with PBS to concentrations of 1 × 10 6 CFU/mL.Each bacterial isolate was injected into 10 G. mellonella, and the survival rates of the G. mellonella were recorded.The experiment was repeated three times.

Statistical analysis
SPSS 27.0 software was used for data processing.The risk factors for CRKP infection and infection-related mortality were analysed by univariate logistic regression analysis.The variables with P < 0.05 in the univariate analysis were included in the multivariate logistic regression model, and backward stepwise regression analysis was performed to identify independent risk factors.A chi-square test was used to analyse the co-occurrence of the four virulence genes in the CRKP and CSKP isolates.P < 0.05 was considered to indicate a statistically significant difference.

Clinical cohort study of mortality: survival versus death in patients with K. pneumoniae infection
Among the 311 patients with K. pneumoniae infection, 57 (18.3%) patients died.Univariate analysis showed that respiratory failure, ICU admission, and prior use of carbapenems were factors related to death of in these patients.Multivariate analysis confirmed that respiratory failure (OR, 5.909; 95% CI, 2.911-11.996;P < 0.001) and ICU admission (OR, 12.285; 95% CI, 4.180-36.109;P < 0.001) were independent risk factors for death in these patients (Table 3).

AST of the K. pneumoniae isolates
All of the K. pneumoniae isolates were resistant to AM.In total, 140 (45.0%)isolates were categorised as CRKP based on their IPM susceptibility.Compared with the CSKP isolates, the CRKP isolates had higher resistance rates to all antibiotics except ampicillin, to which all isolates were resistant.In addition, the CRKP isolates showed resistance rates of 1.4% and 6.4% for polymyxin B and ceftazidime-avibactam, respectively (Table 4).The detailed results of antibiotic susceptibility testing of CSKP and CRKP isolates are presented in Table 4.
The proportional relationships between STs and prior antibiotic use, resistance genes, virulence genes, and resistance ratios to different antibiotics are shown in Fig. 3.

MLST of CRKP
To further investigate this phenomenon, 24 isolates carrying all of the four virulence genes were randomly selected and subjected to the G. mellonella virulence test.It was found that six of these isolates exhibited high virulence with a 72-h G. mellonella survival rate < 50%, 14 isolates showed low virulence with a 72-h G. mellonella survival rate > 50%, and four isolates were found to be nonvirulent with a 72-h G. mellonella survival rate of 100%.The hypervirulent positive-control K. pneumoniae strain NTUH-K2044 was found to have a 36-h G. mellonella survival rate of 0%, whereas seven K. pneumoniae isolates without these four virulence genes and the negative control strain had a 72-h G. mellonella survival rate of 100% (Fig. 4).

Discussion
To improve the efficacy of CRKP infection treatment, we investigated the risk factors for CRKP infection and found ICU admission and prior antibiotic use to be associated with CRKP infection, consistent with previous studies [26].
Hypertension has been reported to be associated with bloodstream infection of CRKP [27], and our study showed that it was also one of the risk factors for CRKP infection.K. pneumoniae and A. baumannii have a mutually beneficial relationship, protecting and feeding each other to promote their survival [28].Plasmids carrying resistance genes can also be transferred between them [29].Accordingly, we found that the coexistence of carbapenem-resistant A. baumannii was associated with CRKP infection.Our results found that patients with neurological disorders were more susceptible to CRKP infection than those without neurological disorders, which may be related to the former patients' need for absolute bed rest.We also found that patients with cancer were more susceptible to CSKP infection than those without cancer, but the reason was not clear and needs to be explored in a further study.
Our results demonstrated that different STs tended to carry different resistance genes.As in many other studies in China, a high clonal diversity of K. pneumoniae was observed in our study, with ST11 being the predominant ST [8,[30][31][32].Another important phenomenon observed was that all of the ST11 isolates were CRKP with a bla KPC-2 positive rate of 93.8%, and ST11 isolates harbouring bla KPC-2 (ST11-KPC-2) accounted for 80.0% of the CRKP isolates and 36.0% of the total K. pneumoniae isolates.This suggests that ST11-KPC-2 was responsible for most of the K. pneumoniae infection cases.Thus, there is an urgent need to prevent and control the further spread of ST11-KPC-2 in hospitals across China.Furthermore, this study was the first to identify ST438, ST751 and ST3984 carrying bla KPC-2 ; ST111 and ST5035 carrying bla NDM like; and ST1049 carrying both bla KPC-2 and bla NDM like.
Infection cases attributable to hv-CRKP isolates are increasing by the year in hospitals, and plasmid recombination and fusion events have been found to be the two important evolutionary pathways responsible for the emergence of hv-CRKP isolates [22,33].The virulence plasmid pLVPK containing two capsular polysaccharide regulator genes (rmpA and rmpA2) and several siderophore gene clusters (e.g.iucABCD/iutA/iroBCDN clusters) is thought to contribute to the hypermucoviscous phenotype of K. pneumoniae [20,18,19].Therefore, we further investigated the presence of four representative virulence genes carried by pLVPK in our isolates.We found that 133 (42.8%) of all K. pneumoniae isolates and 83 (59.3%) of the CRKP isolates carried all of the four virulence genes rmpA, rmpA2, iroN and iucA, Fig. 3 Heatmap of virulence gene carrier rates, multiple antibiotic use rates, carbapenem antibiotic use rates and antimicrobial resistance rates in different MLST CRKP strains and 78 of those 83 (94.0%)CRKP strains belonged to ST11-KPC-2.Such high proportions of K. pneumoniae, CRKP and ST11-KPC-2 isolates carrying all of the four virulence genes have not been previously reported.A study reported that among 1052 CRKP strains isolated during 2015-2017 from 56 centres across China, only 72 (6.8%) strains carried all of the four virulence genes [34].In recent years, the highest prevalence of a single virulence gene in CRKP isolates was approximately 20% in China and Vietnam, followed by 10% in Qatar, and less than 10% in the USA and the UK [21,[35][36][37].Thus, the mechanism underlying the simultaneous presence of all four of these virulence genes at high rates in K. pneumoniae, CRKP and ST11-KPC-2 isolates is currently unknown and needs to be investigated in the future.
In K. pneumoniae, the occurrence of pLVPK or pLVPKlike virulent plasmids was reported to have a strong correlation with highly hypervirulent phenotypes.Several experiments have confirmed the hypervirulent phenotype of pLVPK-positive isolates, including the string test, serum killing assay and the G. mellonella infection model [20,22].Highly hypervirulent phenotypes of K. pneumoniae have been reported to be associated with high mortality rates [38][39][40][41].However, in the present study, carbapenem resistance rather than the co-occurrence of the virulence genes rmpA, rmpA2, iroN and iucA in the isolates was associated with patient death during hospitalisation.The possible reasons for the inconsistency include low expression levels of virulence genes under host immune pressure, prior use of two or more antibiotics or carbapenems, serious underlying diseases and a small sample size in our study.Supporting our speculation, only a small proportion of the K. pneumoniae isolates containing these four virulence genes was found to exhibit high virulence in the G. mellonella infection model.
Given our study's limitations of being a single-centre study and having a small sample size, a multicentre study with a large sample size is warranted to confirm our results.
In conclusion, the independent risk factors for CRKP infection in our hospital were diverse.ST11-KPC-2, which carried all of the four virulence genes, was responsible for most of the K. pneumoniae infection cases.Furthermore, patient mortality was associated with carbapenem resistance rather than with the co-occurrence of all of the four virulence genes.

Fig. 2
Fig. 2 Genotyping characteristics of various drug resistance genes of CRKP

Table 1
Primers used to screen for the presence of bla KPC-2 , bla NDM like and bla OXA-48 like in K. pneumoniae

Table 2
Clinical characteristics of patients infected with K. pneumoniae OR: odds ratio; CI: confidence interval; ICU: intensive care unit.The variables with P < 0.05 screened in univariate analysis were included in multivariate logistic regression analysis

Table 5
Analysis of K.