Prevalence and characteristics of surgical site hypervirulent Klebsiella pneumoniae isolates

Abstract Background We aim to determine the prevalence of hypervirulent Klebsiella pneumoniae (hvKp), which causes surgical site infections (SSIs), and describe the microbiological and molecular characteristics of hvKp isolates. Methods Non‐duplicate K. pneumoniae strains were isolated from wound drainage specimens of postoperative patients at the Chinese PLA General Hospital between September 2008 and July 2017. Antimicrobial susceptibility, string test, pulsed‐field gel electrophoresis (PFGE), and genome sequencing analyses were performed. Results Fifty‐one K. pneumoniae strains were isolated from wound drainage specimens collected from postoperative patients. Twenty‐six hvKp strains, including 17 (17/37, 46.0%) and 9 (9/14, 64.3%) hvKp strains, were isolated from 37 and 14 patients with SSIs and community‐acquired infections (CAIs), respectively. Notably, 4 extended‐spectrum beta‐lactamase (ESBL)‐producing hvKp strains (4/26, 15.4%) and 2 carbapenem‐resistant hvKp strains (2/26, 7.7%) were found. Thirteen K1 serotype (13/26, 50.0%) and 7 K2 serotype (7/26, 26.9%) strains were identified. Phylogenetic analysis results showed that 13 K1 serotype isolates exhibited a high degree of clonality, while 7 K2 serotype strains were genetically unrelated. MLST analysis indicated that there was a strong correlation between ST23 and the K1 serotype. ST65, ST86, and ST375 were prevalent in K2 serotype strains. Almost all hvKp strains (24/26, 92.3%) harbored large virulence plasmids with a high degree of homology to pNTUH‐K2044 and sizes ranging from 140 to 220 kbp. Conclusions HvKp strains were prevalent in SSIs. Effective surveillance and control measures should be implemented to prevent the dissemination of such organisms, including the ESBL‐producing and carbapenem‐resistant hvKp strains.


| INTRODUC TI ON
The new variant of Klebsiella pneumoniae, hypervirulent K. pneumoniae (hvKp), is characterized by its ability to cause life-threatening, metastatic, and invasive infections, such as liver abscesses, pneumonia, meningitis, necrotizing fasciitis, and endophthalmitis in healthy individuals in the community. 1 After it was first described in Taiwan in 1986, 2 the hvKp epidemic has spread in Asia, and has reportedly spread sporadically and at increasing rates in many other countries in North America, Europe, South America, Middle East, and Australia. The hvKp strain is different from classic K. pneumoniae (cKp), as it typically exhibits a hypermucoviscosity phenotype, generally identified using the "string test." Initially, hvKp has been defined by its hypermucoviscosity phenotype 3,4 ; however, it is unclear whether all hvKp strains exhibit hypermucoviscosity. 1 Several putative virulence genes, including the regulator genes of the mucoid phenotype A (rmpA) and siderophore aerobactin (iucABCD), have been found to contribute to hvKp virulence 5,6 and are located in a large virulent plasmid.
Meanwhile, aerobactin, which accounts for >90% of siderophore production, 7 significantly enhances the ex vivo survival of hvKp in human ascites and serum, and enhances in vivo survival in mice models of systemic and pulmonary infection. 8 This suggests that aerobactin could be regarded as a crucial virulence factor and new defining characteristic for hvKp. 9,10 Here, hvKp was defined based on the presence of at least two of the following indicators: a positive string test, rmpA/rmpA2 gene-positive status, and/or aerobactin-positive status.
In contrast to cKp, hvKp is rarely resistant to commonly used antimicrobial agents, with the exception of ampicillin. However, extended-spectrum beta-lactamase (ESBL)-producing and carbapenem-resistant hvKp strains have been observed, especially with healthcare-associated infections (HAIs), in long-term care facilities. [11][12][13] Surgical site infections (SSIs), one of the most frequent types of HAIs, greatly increase the burden on the patient and healthcare system, while the prevalence and molecular characteristics of SSI-causing hvKp isolates remain limited. This study aims to investigate the prevalence of hvKp-causing SSIs and describe the microbiological and molecular characteristics of hvKp isolates.

| Bacterial isolates and antimicrobial susceptibility testing
Fifty-one non-duplicate K. pneumoniae strains were isolated from wound drainage specimens of postoperative patients at the Chinese PLA General Hospital, between September 2008 and July 2017.
Bacterial identification was performed using MALDI-TOF (bio-Mérieux), and results were further confirmed by 16S rRNA gene sequencing. The ESBL production level and antimicrobial susceptibility to ceftazidime, ceftriaxone, cefepime, aztreonam, ertapenem, imipenem, amikacin, gentamicin, ciprofloxacin, levofloxacin, and sulfamethoxazole-trimethoprim were assessed using Vitek II (bioMérieux), and results were interpreted according to the 2018 Clinical and Laboratory Standards Institute (CLSI) guidelines.

| Case definitions and clinical data collection
According to the US Centers for Disease Control and Prevention (CDC), SSI is an infection that occurs within 30 days after the operative procedure and involves at least one of the following: purulent drainage; organisms isolated in wound culture; at least one of the following signs of infection: erythema, heat, pain, and swelling of surgical incision; or diagnosis of a SSI by the surgeon. 14 An abscess observed in a surgical site before the operation should be considered a CAI instead of an SSI. For each case, the following data el-

| String test
The string test was used to identify the hypermucoviscosity phenotype. Positive results were obtained if a viscous string with a length >5mm was formed, by the stretching of bacterial colonies on a blood agar plate after overnight incubation, using an inoculation loop.

Demographic data and clinical characteristics have been shown in
The antimicrobial resistance rates of ceftazidime, ceftriaxone, aztreonam, and gentamicin in patients infected with cKp were significantly higher than those for patients infected with hvKp strains (P < .05), while there were no significant differences in antimicrobial resistance rates of amikacin, cefepime, ertapenem, imipenem, ciprofloxacin, levofloxacin, and sulfamethoxazole/trimethoprim for patients infected with hvKp and cKp (Table 2). Notably, 4 ESBLproducing hvKp strains (4/26, 15.4%) and 2 carbapenem-resistant hvKp strains (2/26, 7.7%) were detected.  Moreover, all the K1 serotype isolates were found to cluster in a single clade and were clearly separated from the other isolates. This revealed the high degree of clonality of the K1 serotype; the 7 K2 serotype strains were found to be genetically unrelated (Figure 1).

| Virulence characteristics of hvKp isolates
Among the 26 hvKp isolates, 24 strains harbored a virulence plasmid with a high homology to pNTUH-K2044 and included 13 K1 serotype, 7 K2 serotype, 3 K64 serotype, and 1 K54 serotype strains ( Figure 2). Moreover, the S1-PFGE and BLASTN atlas illustrated that all the 13 K1 serotype strains contained a virulence plasmid with a size similar to that of pNTUH-K2044 (around 220 kbp), while 7 K2 serotype strains harbored a virulence plasmid with different sizes, ranging from 140 to 220 kbp. Interestingly, 6 strains harbored another 1-3 plasmids, besides a virulence plasmid (Figure 3). One K5 serotype and one K64 serotype strain did not exhibit a virulence plasmid. Both contained the ICEkp1 virulence island in its chromo-

some. The detection rates of virulence genes for individuals infected
with hvKp, including colibactin, aerobactin, salmochelin, rmpA, and rmpA2, were higher than those for individuals infected with cKp (Table 3).  In summary, we have shown that hvKp (defined as aerobactin positive) strains were prevalent in SSI infections in China. K1 and K2

| D ISCUSS I ON
were the major serotypes of hvKp, and almost all the hvKp strains (24/26, 92.3%) harbored a large virulence plasmid with a high homology to pNTUH-K2044. Effective surveillance and control measures should be implemented to prevent the dissemination of such organisms, including the ESBL-producing and carbapenem-resistant hvKp strains.

E TH I C A L A PPROVA L
Not applicable.