The Epidemiology, Virulence and Antimicrobial Resistance of Invasive Klebsiella pneumoniae at a Children’s Medical Center in Eastern China

Introduction

Invasive bacterial infection caused by Klebsiella pneumoniae (K. pneumoniae) is one of the most common diseases in children.^1,2 The infection can involve the respiratory tract, blood, nervous system and surgical site, and lead to pneumonia, bacteremia/septicemia, meningitis and abscesses.² In severe clinical cases, it can also lead to multiple organ failure, or even death.³

In recent years, infection with hypervirulent K. pneumoniae (HVKP) has been reported in many countries and has become a global public health concern.^4–6 The virulence of HVKP is associated with capsular polysaccharides, siderophores and other virulence factors.^7–9 Previous studies of HVKP mainly screened for aerobactin, which is encoded by iucA.⁹ However, more hypervirulence genes such as _prmpA, _rmpA2, _crmpA, peg-344, terB, iroB and irp2 have recently been identified.⁷ These HVKP strains warrant closer attention as they have the ability to cause severe and life-threatening infections in healthy individuals through these virulence factors.⁶

The development of drug resistance by K. pneumoniae and the emergence of carbapenem-resistant K. pneumoniae (CRKP) caused by the misuse and overuse of antibiotics is also a major threat to public health.^10,11 The isolation rate of CRKP is higher among children than in adults and is increasing in China.^12,13 Meanwhile, the multidrug-resistant (MDR) is more likely to occur than CRKP,^14–16 and MDR pathogens can be transmitted to humans via the food chain or through direct contact with infected animals,^17–19 which should be of concern as well.

HVKP and CRKP were not traditionally overlapped.¹³ However, cases of infection with carbapenem-resistant and hypervirulent K. pneumoniae (CR-HVKP) have recently been reported, which have a poor prognosis and are challenging to treat.^20–23 To date, there have been limited studies of CR-HVKP isolated from pediatric patients. On this basis, the present study investigated the virulence and antibiotic resistance characteristics of invasive K. pneumoniae at a children’s medical center in eastern China.

Materials and Methods

Study Site

This study was conducted in Suzhou, a major city with a population of 12.7 million in the southeast area of Jiangsu Province in eastern China. The population of children aged under 15 years old was 1.7 million in 2020. The Children’s Hospital of Soochow University (CHSU), located in the central area of Suzhou, is a children’s medical center in eastern China and the only provincial tertiary children’s hospital in Jiangsu Province. CHSU has 1500 beds and serves >70,000 inpatients and >2 million outpatients annually. This study had no impact on patients and was approved by the Ethics Committee of CHSU (No. 2020CS099).

Bacterial Strains

Nonduplicated invasive K. pneumoniae isolates were collected at CHSU from January 1, 2016 to December 31, 2020. Invasive K. pneumoniae was defined as K. pneumoniae isolated from a normally sterile site, namely blood, cerebrospinal fluid (CSF), pleural fluid or joint fluid. All invasive K. pneumoniae strains were confirmed by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS, Bruker, Mannheim, Germany).²⁴

Detection of Capsular Serotypes and Virulence-Associated Genes

The common hypervirulent capsular serotypes (K1, K2, K5, K20, K54 and K57), hypervirulence genes (iucA, _prmpA, _prmpA2, _crmpA, peg-344, terB, iroB and irp2) and other virulence genes (ybtS, mrkD, entB, kfu, allS, iutA, k₂A, wabG, Uge, fimH, wcaG, Kpn, ycfM, iroN, Hly and cnf-1) of invasive K. pneumoniae isolates were screened by PCR as previously described.^7,13,25 The PCR products were analyzed by 1.5% agarose gel electrophoresis, and positive products were sequenced and aligned with sequences in GenBank using the Basic Local Alignment Search Tool (BLAST).

Antimicrobial Susceptibility Test

The antimicrobial susceptibility of invasive K. pneumoniae isolates was determined by the Kirby–Bauer (K-B) method and Vitek 2 Compact system (bioMérieux, Marcy-l’Etoile, France) according to the manufacturer’s instructions. The results were analyzed according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI).^24,26

Detection of β-Lactam Antibiotic and Colistin Resistance Genes

PCR was performed to detect carbapenemase genes (bla_KPC, bla_OXA, bla_NDM, bla_VIM, bla_IMP, bla_AIM, bla_GIM, bla_SIM, bla_SPM, bla_BIC and bla_DIM), extended-spectrum β-lactamase (ESBL) genes (bla_TEM, bla_SHV, bla_CTX-M, bla_VEB and bla_PER) and AmpC cephalosporinase genes (bla_CMY, bla_ACT, bla_ACC, bla_DHA, bla_FOX, bla_MOX, bla_CIT and bla_EBC).^13,27 Colistin resistance genes (MCR-1 and MCR-2) were also examined in parallel.^28,29 The primers used to screen the abovementioned genes have been previously reported and the PCR products were analyzed as described above.

Multilocus Sequence Typing (MLST)

MLST was performed using the primers and protocol available from the Pasteur MLST website (https://bigsdb.pasteur.fr/klebsiella/primers_used.html). Seven housekeeping genes of invasive K. pneumoniae isolates, namely rpoB, gapA, mdh, pgi, phoE, infB and tonB, were analyzed and the sequence types (STs) were compared with those available in the Pasteur MLST database.

Data Analysis

The chi-squared test was used to analyze the data and p<0.05 was considered statistically significant. Antibiotic resistance data were analyzed with WHONET v5.6 software (WHO Collaborating Centre for Surveillance of Antimicrobial Resistance, Boston, MA, USA).

Results

Clinical Characteristics and Epidemiology of Invasive K. pneumoniae Infection

Over the 5-year study period, a total of 94 non-duplicated invasive K. pneumoniae strains were collected and identified at our hospital. The clinical characteristics of the study population are summarized in Table 1. Of the 94 children, 59 (62.77%) were male. The median age at the time of diagnosis was 24 months (range: 3 days-180 months); at the time of infection, 62 children (65.96%) were aged <5 years and 32 (34.04%) were aged ≥ 5 years. Most of the invasive K. pneumoniae strains were isolated from the department of hematology (52.13%) and neonatology (25.53%). The clinical syndromes recognized in children with invasive K. pneumoniae infection were bacteremia (89.36%, 84/94), pneumonia (8.51%, 8/94) and meningitis (2.13%, 2/94). Based on the first isolates, the primary sites of infection were blood (73.40%, 69/94), bronchoalveolar lavage fluid (13.83%, 13/94), pleural fluid (7.45%, 7/94), cerebrospinal fluid (2.13%, 2/94), seroperitoneum (2.13%, 2/94) and urine (1.06%, 1/94) (Figure 1). The annual trends of invasive K. pneumoniae cases by clinical diagnosis are shown in Figure 2. There was an increase in the overall number of cases between 2016 and 2019, primarily driven by an increase in K. pneumoniae bacteremia although there were also marginal increases in the number of cases of invasive K. pneumoniae pneumonia and meningitis. Additionally, 12 strains of invasive K. pneumoniae were CR-HVKP.

Table 1 Clinical Characteristics and Epidemiology of Invasive K. pneumoniae Infection at Children’s Hospital of Soochow University from 2016 to 2020

Figure 1 The first isolated samples of invasive K. pneumoniae strains in this study.

Figure 2 The annual trends of invasive K. pneumoniae cases by clinical diagnosis from 2016 to 2020 in this study.

Distribution of Capsular Types and Virulence-Associated Genes

Previous studies have shown that some capsular serotypes of K. pneumoniae (eg, K1, K2, K5, K20, K54, and K57) contribute to hypervirulence.¹³ In this study, common hypervirulent capsular serotypes were detected in 87 strains (92.55%) of invasive K. pneumoniae; the main capsular serotype was K57 (62.77%, 59/94) and the remaining ones were K2 (11.70%, 11/94), K1 (7.45%, 7/94), K54 (6.38%, 6/94), K5 (4.26%, 4/94) and others (7.45%, 7/94). The K20 capsular serotype was not detected (Table 2 and Figure 3). All invasive K. pneumoniae strains carried virulence-associated genes; among them, 84 (89.36%) carried hypervirulence genes, which were iroB (86.17%, 81/94), peg-344 (34.04%, 32/94), _prmpA (30.85%, 29/94), iucA (13.83%, 13/94), _crmpA (6.38%, 6/94) and _prmpA2 (2.13%, 2/94). The other main virulence genes were wabG (100%, 94/94), mrkD (98.94%, 93/94), ycfM (96.81%, 91/94), fimH (95.74%, 90/94) and Uge (88.30%, 83/94) (Table 2 and Figure 3). There were no significant differences between CR-HVKP and non CR-HVKP strains in terms of capsular types and virulence-associated genes. These data suggest that the isolated invasive K. pneumoniae strains, including CR-HVKP and non CR-HVKP, all carried a high proportion of virulence genes that may contribute to disease.

Table 2 The Capsular Serotypes and Virulence-Associated Genes of Invasive K. pneumoniae Strains in This Study

Figure 3 The heat-map of capsular serotypes and virulence-associated genes of invasive K. pneumoniae strains in this study.

Antimicrobial Susceptibility

The invasive K. pneumoniae strains exhibited different patterns of resistance to various antibiotics, but in general, they showed low resistance to aminoglycosides (amikacin and tobramycin), carbapenems (meropenem, imipenem and ertapenem), cephalosporins (ceftazidime, cefepime and ceftizoxime), cephalosporins+β-lactamase inhibitors (cefoperazone/sulbactam), cephamycins (cefoxitin and cefotetan), fluoroquinolones (ciprofloxacin and levofloxacin), monobactams (aztreonam), penicillins+β-lactamase inhibitors (piperacillin/tazobactam) and penicillins (piperacillin) (Table 3 and Figure 4). Notably, 29 strains (30.85%) of non CR-HVKP produced ESBLs. The resistance rates of CR-HVKP strains were much higher than those of non–CR-HVKP strains (Table 3 and Figure 4). Moreover, there were significant differences (p<0.05) between CR-HVKP and non CR-HVKP strains in terms of susceptibility to aminoglycosides (amikacin), carbapenems, cephalosporins, cephalosporins+β-lactamase inhibitors, cephamycins, penicillins+β-lactamase inhibitors, and penicillins (Table 3 and Figure 4).

Table 3 The Antimicrobial Resistance of Invasive K. pneumoniae Strains in This Study

Figure 4 The heat-map of the different degrees of antimicrobial resistance of invasive K. pneumoniae strains in this study.

Distribution of β-Lactam Antibiotic and Colistin Resistance Genes

Carbapenemase, ESBL and AmpC cephalosporinase genes are the main β-lactam antibiotic resistance genes in Enterobacteriaceae.^13,27 In this study, all invasive K. pneumoniae strains carried β-lactam antibiotic resistance genes. The carbapenemase genes were bla_NDM-1 (19.15%, 18/94), bla_VIM-1(13.83%, 13/94), bla_KPC-2 (10.64%, 10/94) and bla_OXA-1 (5.32%, 5/94); the ESBL genes were bla_SHV-11 (86.17%, 81/94), bla_CTX-M-14 (48.94%, 46/94), bla_TEM-1 (19.15%, 18/94) and bla_VEB-1 (6.38%, 6/94); the AmpC cephalosporinase genes were bla_FOX-1 (86.17%, 81/94), bla_ACT-1 (70.21%, 66/94) and bla_DHA-1 (31.91%, 30/94). Colistin resistance genes (MCR-1 and MCR-2) were not detected (Table 4 and Figure 5). There were no significant differences between CR-HVKP and non CR-HVKP strains in terms of the rates of carbapenemase, ESBL and AmpC cephalosporinase genes (with the exception of bla_ACT-1). These data suggest that invasive K. pneumoniae strains including CR-HVKP and non CR-HVKP harbor a high proportion of β-lactam antibiotic resistance genes simultaneously.

Table 4 The β-Lactam Antibiotic and Colistin Resistance Genes of Invasive K. pneumoniae Strains in This Study

Figure 5 The heat-map of β-lactam antibiotic genes of invasive K. pneumoniae strains in this study.

Characteristics of CR-HVKP Strains

A total of 12 invasive K. pneumoniae strains were identified as CR-HVKP, which carried at least 3 β-lactam antibiotic resistance genes. Meanwhile, all 12 CR-HVKP strains carried hypervirulence and other virulence genes, and had hypervirulent capsular serotypes. Additionally, 7 distinct STs were observed among these CR-HVKP strains: ST29 (25.00%, 3/12), ST14 (16.67%, 2/12), ST17 (16.67%, 2/12), ST37 (16.67%, 2/12), ST45 (8.33%, 1/12), ST101 (8.33%, 1/12) and ST234 (8.33%, 1/12) (Table 5). The 12 CR-HVKP strains had low resistance to gentamicin, amikacin, tobramycin, levofloxacin and aztreonam, but were resistant to most other antimicrobial drugs (Table 3 and Figure 4).

Table 5 The Capsular Serotypes, Genotypes and MLST of CR-HVKP Strains Identified in This Study (n=12)

Correlation Between Phenotypic and Genotypic MDR Patterns in Invasive K. pneumoniae

Besides the 12 CR-HVKP strains, other MDR strains of invasive K. pneumoniae were also observed in this study. A total of 14 strains (14.89%) were MDR (defined as resistant to ≥ 1 agent in ≥ 3 antimicrobial classes³⁰); these strains exhibited MDR to 7 antimicrobial classes (n=3), 6 antimicrobial classes (n=2), 5 antimicrobial classes (n=6), 4 antimicrobial classes (n=2), and 3 antimicrobial classes (n=1). Additionally, 8 of the strains harbored 5 β-lactam antibiotic resistance genes, 5 strains had 4 of these genes, and 1 strain had 3 of the genes (Table 6).

Table 6 The Correlation Between Phenotypic and Genotypic Resistance Patterns Among the MDR K. pneumoniae Strains in This Study (n=14)

Discussion

K. pneumoniae is one of the most common pathogens responsible for nosocomial and community-acquired infections in children, which is associated with high morbidity and mortality.^1–3 In recent years, the increased prevalence of HVKP has become a serious global threat to public health.⁴ HVKP can cause many types of invasive infection, such as severe pneumonia, bacteremia/septicemia and meningitis, which are often associated with severe symptoms and high mortality rates.^4–6 The capsular serotype of K. pneumoniae is one of the factors responsible for its hypervirulence, protecting the bacteria from phagocytosis. There are at least 78 capsular serotypes, with K1, K2, K5, K20, K54 and K57 being the most common hypervirulent serotypes.^8,31 In addition, K. pneumoniae also carries several genes that were shown to contribute to hypervirulence, such as iucA (involved in aerobactin siderophore biosynthesis), the plasmid-borne rmpA gene (_prmpA), _prmpA2, chromosomal gene rmpA (_crmpA) (which regulates of the mucoid phenotype via increased capsule production), peg-344 (a putative transporter), terB (involved in tellurite resistance), iroB (involved in salmochelin siderophore biosynthesis), and irp2 (involved in yersiniabactin siderophore biosynthesis).⁷ K1 and K2 are the most common hypervirulent serotypes of K. pneumoniae.^13,31,32 But in our study, K57 was the main capsular serotype of invasive K. pneumoniae strains, which may be explained by differences in serotype distribution according to geographic areas and across populations. Moreover, all invasive K. pneumoniae strains carried virulence-associated genes and the proportion with hypervirulence genes was significantly higher than that reported in previous studies, indicating that these factors play an important role in the pathogenicity of invasive K. pneumoniae in children.

The prevalence of CRKP is also a global public health problem and the clinical isolation rate of CRKP has been increasing in recent years.^10–13 In China, the rate of detection of CRKP in children increased from 3% to > 20% according to a surveillance program conducted from 2005 to 2017. Notably, this rate was higher than that recorded in adults.^11,13,22 CRKP is resistant to most antimicrobial drugs and there are currently limited therapeutic options available, making infections difficult to control and resulting in high mortality rates.^12,13 The main mechanism of CRKP resistance is the production of enzymes, such as carbapenemases, ESBLs and AmpC, that hydrolyze antimicrobial drugs.^13,27 bla_KPC-2 is the most common carbapenemase in adults in China,^13,33 while the predominant carbapenemases in children were bla_IMP-4, bla_NDM-1 and bla_OXA-232 in different periods.^13,34 In our study, carbapenemase genes were detected in only a few invasive K. pneumoniae strains; however, most strains harbored ESBL and AmpC enzyme genes, mainly bla_SHV-11, bla_FOX-1 and bla_ACT-1. This diverges from previous findings that bla_TEM-1 was the main ESBL gene while the proportion of AmpC genes was low.^27,35 An evaluation of drug susceptibility showed that the 94 invasive K. pneumoniae strains identified in this study had low drug resistance rates but high carriage rates of ESBL and AmpC enzyme genes. Moreover, we found 14 MDR K. pneumoniae strains that carried at least 3 β-lactam antibiotic resistance genes, representing potential threats. Therefore, the drug resistance of invasive K. pneumoniae in children warrants further attention.

In this study, 12 CR-HVKP strains were identified belonging to 7 STs. ST29 was the most common ST in these strains, whereas ST11 was previously reported as the predominant ST among CRKP isolated from both adults and children in China.^27,36 The predominant ST of CRKP and CR-HVKP may differ. However, there have been limited studies of CR-HVKP isolated from pediatric patients and further investigations are required. CR-HVKP has high pathogenicity in children and is resistant to most antimicrobial drugs, which presents a challenge for the clinical management of infections. More research is needed to develop strategies to overcome these challenges.

Conclusions

In summary, our findings reveal the epidemiology, virulence, and antimicrobial resistance of invasive K. pneumoniae at a children’s medical center in eastern China. Both CR-HVKP and MDR strains were identified. HVKP and CRKP were found to overlap in pediatric patients; their accurate identification can aid diagnosis and effective treatment.