Clinical and Microbiologic Analysis of Klebsiella pneumoniae Infection: Hypermucoviscosity, Virulence Factor, Genotype, and Antimicrobial Susceptibility

Hypervirulent Klebsiella pneumoniae (KP) is defined according to hypermucoviscosity or various virulence factors and is clinically associated with community-acquired liver abscess (CLA). In this study, we investigated the clinical and microbiological characteristics of KP and significant factors associated with hypervirulence. The clinical characteristics, antimicrobial susceptibility, hypermucoviscosity, serotypes, hypervirulence-related genes, and biofilm formation of 414 KP isolates collected from the Keimyung University Dongsan Hospital between December 2013 and November 2015 were analyzed according to CLA. Significant risk factors for hypervirulent KP (HvKP) associated with CLA were investigated using logistic regression analysis. Notably, 155 (37.4%) isolates were hypermucoviscous, and 170 (41.1%) harbored aerobactin. CLA was present in 34 cases (8.2%). Epidemiology and treatment outcomes did not differ significantly between the CLA and non-CLA groups. The CLA group had significantly higher antibiotic susceptibility, K1/K2, rmpA, magA, allS, kfu, iutA, string test-positive result, and biofilm mass. Multivariate logistic regression revealed rmpA (OR, 5.67; 95% CI, 2.09–15.33; p = 0.001), magA (OR, 2.34; 95% CI, 1.01–5.40; p = 0.047), and biofilm mass >0.80 (OR, 2.13; 95% CI, 1.00–4.56; p = 0.050) as significant risk factors for CLA. rmpA was identified as the most significant risk factor for CLA among KP strains, implying that it is an important factor associated with HvKP.


Introduction
Klebsiella pneumoniae is a Gram-negative bacterium associated with an invasive syndrome that has caused liver abscesses in Southeast Asian populations over the past three decades [1][2][3].Hypervirulent K. pneumoniae was first described in a patient with a liver abscess in Taiwan in the 1980s [4].At that time, hypervirulence was defined based on hypermucoviscosity, which was confirmed by a string test, a phenotypic screening marker [5,6].When hypervirulent K. pneumoniae first presented, it had the tendency to display low antimicrobial resistance, was associated with community-acquired infections, and was associated with the virulence factor rmpA and biofilm formation [4,[7][8][9].Hypervirulent K. pneumoniae has distinctive features compared with classic K. pneumoniae.Characteristics of classical K. pneumoniae infection were hospital-acquired infections, including pneumonia and urinary tract infections, and higher antimicrobial resistance rates [10,11].
Many studies on hypervirulent K. pneumoniae have been conducted over the years, and some researchers suggested that not only hypermucoviscosity but also aerobactin could be used to evaluate hypervirulence [12][13][14].Aerobactin is a representative siderophore, and its production plays an important role in determining the hypervirulence of K. pneumoniae [15].treatment outcomes of the patients.Acute kidney injury was defined as an increase in serum creatinine level by >0.3 mg/dL within 48 h, an increase in serum creatinine level to >1.5 times baseline within 7 days, or a decrease in urine volume to <0.5 mL/kg/h over 6 h.The McCabe-Jackson score was used as the criterion to predict the survival of patients based on the prognosis of the underlying diseases, which were classified as rapidly fatal, ultimately fatal, and nonfatal [31].The early treatment outcome was determined after 72 h of empirical antibiotic treatment.Early treatment outcomes were classified as complete response, partial response, or treatment failure.A complete response was defined as an improvement in both clinical conditions, and laboratory findings such as level of white blood cell or C-reactive protein.A partial response was defined as either an improved clinical condition or improved laboratory findings.Treatment failure was defined as the worsening of both clinical conditions and laboratory findings.Death of a patient due to K. pneumoniae infection or a complication of the infection within 30 days was defined as an infection-related 30-day mortality.We compared the strains based on their hypermucoviscous phenotype and hypervirulence-associated genes.

Definitions of Hypervirulent K. pneumoniae
In this study, we defined hypervirulent K. pneumoniae as CLA from a clinical perspective.We analyzed the clinical and microbiological characteristics according to the CLA to determine which genotypes and phenotypes were most associated with the CLA group.In addition, as reference data, the results of a comparative analysis between the two definitions based on hypermucoviscosity and aerobactin positivity, which were previously known as hypervirulent K. pneumoniae definitions, were presented in the supplementary tables.A previous study showed that the capsular serotype tended to have high specificity but low accuracy and sensitivity, and because the factors affecting K. pneumoniae hypervirulence may have a complex effect on each other, the capsular serotype was excluded during multivariate analysis [14].

Detection of K. pneumoniae Hypermucoviscosity
The string test was performed to determine the hypermucoviscous phenotype [32].The string test was positive when a bacteriologic inoculation loop could generate a viscous string >5 mm in length by stretching bacterial colonies on an agar plate (Figure 1).

Study Design
Medical records were retrospectively analyzed to identify underlying diseases, predisposing factors, antibiotics used within the last 3 months, previous hospitalization, antimicrobial susceptibility, clinical presentations, currently administered antibiotics, and treatment outcomes of the patients.Acute kidney injury was defined as an increase in serum creatinine level by >0.3 mg/dL within 48 h, an increase in serum creatinine level to >1.5 times baseline within 7 days, or a decrease in urine volume to <0.5 mL/kg/h over 6 h.
The McCabe-Jackson score was used as the criterion to predict the survival of patients based on the prognosis of the underlying diseases, which were classified as rapidly fatal, ultimately fatal, and nonfatal [31].The early treatment outcome was determined after 72 h of empirical antibiotic treatment.Early treatment outcomes were classified as complete response, partial response, or treatment failure.A complete response was defined as an improvement in both clinical conditions, and laboratory findings such as level of white blood cell or C-reactive protein.A partial response was defined as either an improved clinical condition or improved laboratory findings.Treatment failure was defined as the worsening of both clinical conditions and laboratory findings.Death of a patient due to K. pneumoniae infection or a complication of the infection within 30 days was defined as an infection-related 30-day mortality.We compared the strains based on their hypermucoviscous phenotype and hypervirulence-associated genes.

Definitions of Hypervirulent K. pneumoniae
In this study, we defined hypervirulent K. pneumoniae as CLA from a clinical perspective.We analyzed the clinical and microbiological characteristics according to the CLA to determine which genotypes and phenotypes were most associated with the CLA group.In addition, as reference data, the results of a comparative analysis between the two definitions based on hypermucoviscosity and aerobactin positivity, which were previously known as hypervirulent K. pneumoniae definitions, were presented in the supplementary tables.A previous study showed that the capsular serotype tended to have high specificity but low accuracy and sensitivity, and because the factors affecting K. pneumoniae hypervirulence may have a complex effect on each other, the capsular serotype was excluded during multivariate analysis [14].

Detection of K. pneumoniae Hypermucoviscosity
The string test was performed to determine the hypermucoviscous phenotype [32].The string test was positive when a bacteriologic inoculation loop could generate a viscous string >5 mm in length by stretching bacterial colonies on an agar plate (Figure 1).

Polymerase Chain Reaction
Capsular serotypes and virulence factors, including rmpA (461), magA (1283), allS (764), mrkD (340), ybtS (242), kfu (638), and iutA (920), were identified using multiplex polymerase chain reaction (PCR).Strains were serotyped as K1, K2, K5, K20, K54, or K57 or as non-determined when a specific serotype could not be identified.The primers for bla SHV-1a were used for the positive control reactions.Amplification was performed using a C1000 Thermal Cycler (Bio-Rad, Pleasanton, CA, USA).Crude DNA was prepared by the lysis of the colonies at 100 • C for 10 min in 500 mL of sterile distilled water, followed by centrifugation.The lysed supernatant was used for PCR.The PCR program comprised an initial activation step at 95 • C for 15 min, followed by 30 cycles of 94 • C for 30 s, 60 • C for 90 s, 72 • C for 60 s, and a final extension period at 72 • C for 10 min.The amplicons were separated via electrophoresis at 100 V for 2 h using a 2% agarose gel.Specific primers used to detect the alleles of the target gene sequences are listed in Table S1.

Biofilm Formation
Biofilm mass was determined using a microtiter plate assay and based on optical density (OD, 570 nm).To measure biofilm formation, the bacterial suspensions were incubated in 96-well plates.After overnight incubation at 37 • C for 24 h, crystal violet was added, thereby staining the biofilm purple.Biofilm mass was then measured using the corresponding OD 570 nm of the supernatant following solubilization of crystal violet in 99% ethanol.We used Synergy/HTX Spectrophotometer (BioTek instrument, Inc.Winooski, VT, USA) for analysis of biofilm.Each strain was tested in triplicate, with a positive control of K. pneumoniae ATCC 700603 and a negative control of K. pneumoniae ATCC 13883.

Antimicrobial Susceptibility Test
The isolates were analyzed using an automated microbial identification (Vitek2 Gramnegative identification system; bioMerieux, Lyon, France) and susceptibility test system (Vitek2 AST-N224 system).Antimicrobial susceptibility profiles were determined based on the breakpoints recommended in the guidelines of the 2012 Clinical and Laboratory Standards Institute (CLSI).Extended-spectrum β-lactamase (ESBL) production was detected using an automated methodology based on the Vitek2 AST-N224 system, which was used to inoculate and incubate bacteria according to the manufacturer's recommendations.

Statistical Analyses
Statistical analyses were performed using Statistical Package for the Social Sciences version 21.0 (IBM Corp., Armonk, NY, USA).Categorical variables were compared using the chi-squared test or Fisher's exact test.For continuous variables, the normality of distribution was evaluated using the Kolmogorov-Smirnov test.The Mann-Whitney U test and independent t-test were performed for data that followed non-normal and normal distributions, respectively.Statistical significance was defined as p < 0.05.Risk factors for the CLA group were assessed and analyzed using logistic regression in the total group.An independent variable with p < 0.1 in the univariate analysis was included in the multivariate analysis, and a variable with a final p < 0.05 was considered a significant risk factor.

Clinical and Microbiological Characteristics According to Hypermucoviscosity
In total, 414 K. pneumoniae isolates were included in this study (Figure 2).A hypermucoviscous phenotype was observed in 155 isolates (37.4%).Male sex was more strongly associated with the hypermucoviscous phenotype (69.0%) than with the stringnegative group (54.4%) (p = 0.003).In underlying diseases, solid tumors were less associated with hypermucoviscous phenotype (p = 0.002).Based on the McCabe classification, ultimately, fatal disease was less frequently associated with the hypermucoviscous phenotype (p = 0.017).In predisposing factors, percutaneous catheter drainage (p = 0.006) and invasive procedures (p = 0.001) were associated with the hypermucoviscous phenotype.
In the infection category, the proportion of community-acquired infection was more associated with hypermucoviscous phenotype (p < 0.001).Nosocomial infection was more associated with the string-negative group (p < 0.001).Among infection sources, intraabdominal infection was more strongly associated with the hypermucoviscous phenotype than with the string-negative group (p = 0.001), whereas urinary tract infection showed a weaker association (p < 0.001).Regarding clinical presentations, metastatic infection, severe sepsis and septic shock, concomitant bacteremia, and admission to the intensive care unit (ICU), no differences were observed between the two groups.There were also no significant differences in the treatment outcomes, treatment failure, infection-related 30-day mortality, or acute kidney injury (Table S1).
Diagnostics 2024, 14, x FOR PEER REVIEW 5 of 14 fatal disease was less frequently associated with the hypermucoviscous phenotype (p = 0.017).In predisposing factors, percutaneous catheter drainage (p = 0.006) and invasive procedures (p = 0.001) were associated with the hypermucoviscous phenotype.In the infection category, the proportion of community-acquired infection was more associated with hypermucoviscous phenotype (p < 0.001).Nosocomial infection was more associated with the string-negative group (p < 0.001).Among infection sources, intra-abdominal infection was more strongly associated with the hypermucoviscous phenotype than with the string-negative group (p = 0.001), whereas urinary tract infection showed a weaker association (p < 0.001).Regarding clinical presentations, metastatic infection, severe sepsis and septic shock, concomitant bacteremia, and admission to the intensive care unit (ICU), no differences were observed between the two groups.There were also no significant differences in the treatment outcomes, treatment failure, infection-related 30day mortality, or acute kidney injury (Table S1).
The proportion of community-acquired infection was more associated with the aerobactin-positive group (p < 0.001).Nosocomial infection was more associated with the aerobactin-negative group (p < 0.001).Regarding the infection source, intra-abdominal infection was more strongly associated with the aerobactin-positive group (p = 0.023), whereas urinary tract infection showed a weaker association (p < 0.001).In terms of clinical presentations, metastatic infection, concomitant bacteremia, and admission to ICU, no differences were observed between the two groups.Similarly, no significant differences were observed in the treatment outcomes, treatment failure, infection-related 30-day mortality, or acute kidney injury (Table S3).

Epidemiology and Clinical characteristics According to Community-Acquired Liver Abscess
CLA was detected in 34 isolates (8.2%).In epidemiology, there was no significant difference in male sex and age between the CLA and non-CLA groups.In underlying diseases, solid tumors (p = 0.022) and neurological disease (p = 0.012) were more associated with the non-CLA group (p = 0.022).Based on the McCabe classification, ultimately, fatal disease was less frequently associated with the CLA group (p = 0.004).In predisposing factors, percutaneous catheter drainage (p < 0.001), L-tube (p = 0.015), and invasive procedures (p < 0.001) were associated with the CLA group.In others, such as recent operation (p = 0.019) and cases of prior ICU admission within 1 month (p = 0.014) were associated with the non-CLA group.
In the category of infection, all infections in the CLA group were community-acquired infections, as defined.Regarding the infection source, all infections in the CLA group were intra-abdominal infections, as defined.In terms of clinical presentations, metastatic infection, severe sepsis and septic shock, and admission to ICU, no differences were observed between the two groups.The rate of concomitant bacteremia was significantly higher in the CLA group than in the non-CLA group (61.8% vs. 31.8%;p < 0.001).There were no significant differences in treatment outcomes, treatment failure, infection-related 30-day mortality, or acute kidney injury (Table 1).

Antimicrobial Susceptibility and Microbiological Characteristics According to Community-Acquired Liver Abscess
Regarding antimicrobial susceptibility, most antibiotics such as amoxicillin/clavulanate, aztreonam, cefazolin, cefepime, cefotaxime, ceftazidime, ciprofloxacin, piperacillin/tazobactam, and trimethoprim/sulfamethoxazole had significantly lower resistance to K. pneumonia in the CLA group than in the non-CLA group.ESBL positivity was lower in the CLA group than in the non-CLA group (p < 0.001) (Table 2).

Discussion
In this study, we found that the proportion of hypermucoviscosity and aerobactin gene expression of K. pneumoniae varies widely.The proportion of hypervirulence differed depending on the used definitions of hypervirulent K. pneumoniae whether hypermucoviscosity and aerobactin positivity or not.Significant factors in CLA, for which the definition of hypervirulent K. pneumoniae was first proposed, were rmpA, magA, and biofilm mass.
When comparing the two groups according to the string test, the characteristics of hypervirulent K. pneumoniae were similar to those already known.As with other study results, hypermucoviscosity and aerobactin positivity were not consistent [33].The hypermucoviscous capsule of hypervirulent K. pneumoniae is a key factor of hypervirulence [34].Hypermucoviscous capsule contributes to reduced human cell binding and evasion of neutrophil-mediated phagocytosis.Because of this mechanism, the overproduction of hypermucoviscous capsules by hypervirulent K. pneumoniae has been reported as an important factor that aids bacterial dissemination and metastatic infections in the host [35].Capsules could be affected by various environments and conditions [36].Several studies have shown that the string test is not an appropriate method for assessing K. pneumoniae hypervirulence [13,15,23,37].
When comparing the two groups according to aerobactin results, the characteristics of hypervirulent K. pneumoniae were similar to already known [13].Hypervirulent K. pneumoniae is associated with the possession of large virulence plasmids [37].Iron acquisition, increased capsule production, K1/K2 capsular serotypes, and the colibactin toxin have been identified as the four microbiological and genotypic characteristics of hypervirulent K. pneumoniae [37].Among various microbiologic factors, the ability to acquire iron is essential for bacterial growth [38].High-affinity iron uptake systems contribute to the virulence of K. pneumoniae [39].Aerobactin is located on a large virulence plasmid of K. pneumoniae that is not present in most classic K. pneumoniae strains [40].Aerobactin mediates the virulence of K. pneumoniae and accounts for increased siderophore production under iron-limiting conditions by hypervirulent K. pneumoniae [38].Therefore, some researchers have used aerobactin-positive strains to define hypervirulent K. pneumoniae.However, aerobactin-positive strains do not completely correlate with the existing hypermucoviscous phenotype [12,13,16,41].
Owing to changes in the medical environment, findings that deviate from the previously reported characteristics of hypervirulent K. pneumoniae, such as an increase in antimicrobial resistance and the relationship between healthcare-associated infections, have been reported [15,21,42].In cases of our study when only aerobactin was confirmed positive, and the string test was negative, the antimicrobial resistance rate tended to be higher compared to aerobactin-positive cases; thus, it can be assumed that the characteristics of the hypermucoviscous phenotype may be lost upon exposure to antibiotics [43].In a Chinese study, hypervirulence was defined as a positive result for aerobactin, and approximately 75% of hypervirulent and 18% of classic K. pneumoniae strains were found to be string testpositive [13].In another Chinese study, hypervirulent K. pneumoniae was defined as having both a hypermucoviscous phenotype and a positive result for aerobactin.In that study, the prevalences of rmpA and magA were found to be 81.3% and 78.8% in hypervirulent K. pneumoniae and 17.9% and 61.1% in classic K. pneumoniae, respectively [16].
Hypervirulent K. pneumoniae was distinguished from classic K. pneumoniae based on the clinical presentation of a community-acquired pyogenic liver abscess with metastatic infections, such as endophthalmitis, central nervous system involvement, lung involvement, and antimicrobial susceptible pathogen.[3,44,45].The capsular serotype, determined by surface antigens, has also been reported as an important factor that may affect the virulence of K. pneumoniae [46].Capsular serotypes of K1 and K2 in liver abscess and biofilm formation were more frequently associated with hypervirulent K. pneumoniae [7,[47][48][49].In pyogenic liver abscesses of K. pneumoniae, aerobactin was more correlated than hypermucoviscosity [50].K1 and K2 are the predominant capsular serotypes of hypervirulent K. pneumoniae [27,51], with K1 being the most common capsular serotype, followed by K2 [9,52].In several studies, K1 and K2 showed different expression of virulence factors, especially rmpA, magA, and aerobactin, and displayed higher levels of biofilm formation than other capsular serotypes [25,26,53].The K1 serotype is associated with siderophore iron acquisition systems and invasive infections [18], whereas K2 has a higher diversity of sequence types [26,54,55].
Hypermucoviscosity is still being used as a criterion for hypervirulent K. pneumoniae in many studies [5,21,32,38].Among the virulence factors, rmpA is particularly known to affect capsule production in hypervirulent K. pneumoniae [56].Hypervirulent K. pneumoniae with mutations in rmpA lose hypermucoviscous phenotype and show a strong reduction in virulence [57].Proportions of hypermucoviscosity and rmpA in the aerobactin-positive cases were similar to those of another study, and the distribution of other virulence factors, such as magA, was confirmed to be diverse [58].Hypermucoviscosity, rmpA, aerobactin, and serotype K1 are useful laboratory markers when suspecting community-acquired K. pneumoniae bacteremic liver abscess [59].In cases of K. pneumoniae liver abscess in China, all strains were rmpA-positive, and two-thirds of strains were magA-positive [60].rmpA is specifically correlated with abscess formation in hypermucoviscous K. pneumoniae strains [61].
This study had several limitations.First, it was conducted retrospectively in a tertiary hospital, which may introduce bias in the data interpretation.Second, this study was conducted at a single center.These factors can significantly influence the results and their generalizability.Third, the patients admitted to a tertiary hospital may exhibit more severe symptoms than those in a primary medical center.As patients who were transferred to other hospitals were excluded, there was a limitation in determining the overall condition of patients with K. pneumoniae infections who visited our hospital.Finally, the number of analyzed strains was relatively low.
Despite these limitations, our findings determined that both definitions were useful, as they showed clinical and microbiological features suggestive of hypervirulent K. pneumoniae and that the virulence gene rmpA was significantly higher in CLA with hypervirulent K. pneumoniae.Additional discussion is needed in the future regarding the definition of hypervirulent K. pneumoniae.

Supplementary Materials:
The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/diagnostics14080792/s1.Table S1: Baseline characteristics and clinical presentations of Klebsiella pneumoniae isolates according to hypermucoviscosity; Table S2: Antimicrobial resistance and microbiological characteristics of Klebsiella pneumoniae isolates according to hypermucoviscosity; Table S3: Baseline characteristics and clinical presentations of Klebsiella pneumoniae isolates according to aerobactin positivity; Table S4: Antimicrobial resistance and microbiological characteristics of Klebsiella pneumoniae isolates according to aerobactin positivity.Informed Consent Statement: Patient consent was waived due to the retrospective nature of the study and the use of identifiable specimens.

Data Availability Statement:
The dataset of the current study is available from the corresponding author upon request.

Figure 1 .
Figure 1.(A) Colonies are circular, convex with entire margin, and mucoid.(B) Stretching of the K. pneumoniae colonies resulted in the formation of a string ≥5 mm in length, demonstrating the hypermucoviscous phenotype.

Figure 1 .
Figure 1.(A) Colonies are circular, convex with entire margin, and mucoid.(B) Stretching of the K. pneumoniae colonies resulted in the formation of a string ≥5 mm in length, demonstrating the hypermucoviscous phenotype.

Figure 2 .
Figure 2. Distributions of the hypermucoviscosity, aerobactin positivity, and community-acquired liver abscess of K. pneumoniae in this study group.

Figure 2 .
Figure 2. Distributions of the hypermucoviscosity, aerobactin positivity, and community-acquired liver abscess of K. pneumoniae in this study group.

Author Contributions:
Conceptualization, data curation, laboratory experiment, and methodology, M.H. and H.A.K.; formal analysis, M.H. and J.Y.L.; writing-review and editing, M.H., J.Y.L. and H.A.K.All authors have read and agreed to the published version of the manuscript.Funding: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIP) (NRF-2020R1F1A1070470).Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Keimyung University Dongsan Medical Center (IRB no.2022-05-040).

Table 1 .
Baseline characteristics and clinical presentations of Klebsiella pneumoniae isolates according to CLA.
Values are presented as mean ± standard deviation or number (%).* Fisher's exact test.CLA: community-acquired liver abscess; ICU: intensive care unit.

Table 2 .
Antimicrobial resistance and microbiological characteristics of Klebsiella pneumoniae isolates according to CLA.

Table 3 .
Significant virulence factors for community-acquired liver abscess in Klebsiella pneumoniae using logistic regression analysis.