Clinical and microbiological characteristics and occurrence of Klebsiella pneumoniae infection in Japan.

PURPOSE: Klebsiella pneumoniae is a pathogen that causes pneumonia and urinary tract infection. Hypervirulent K. pneumoniae strains often show hypermucoviscosity, are of the K1 or K2 serotype, and harbor the rmpA and magA genes. However, the differences in the prevalence of K. pneumoniae with these hypervirulent characteristics between the infection and colonization status are not well understood. Therefore, in this study, we compared the clinical and microbiological characteristics of K. pneumoniae isolated from urine or sputum samples of cases of infection and colonization. PATIENTS AND METHODS: This retrospective study was conducted at a tertiary care teaching hospital in Tokyo, Japan. Patients whose sputum or urine tested positive for the presence of K. pneumoniae isolates were randomly included in the study. Clinical and microbiological data were collected from medical records.
RESULTS: Of the 130 cases investigated, 68 and 62 cases showed the presence of K. pneumoniae in the sputum and urine, respectively. There were 49 infection cases, including 21 in the sputum group and 28 in the urine group. The infections were not accompanied by liver abscess. Of the 130 K. pneumoniae isolates, 25 (19.2%) showed capsular serotype K1 or K2, whereas 33 (25.4%) showed hypermucoviscosity. The prevalence of virulence genes magA, allS, rmpA, mrkD, uge, kfu-BC, and wabG was 10% (all in K1), 13.1%, 16.9%, 85.4%, 79.2%, 36.9%, and 91.5%, respectively. In both the sputum and urine groups, there was no difference in the characteristics of patients with infection and those with colonization. Analysis of microbiological characteristics revealed that only rmpA was significantly more frequent in the infection cases than in the colonization/asymptomatic cases in both the sputum and urine groups.
CONCLUSION: The rmpA-positive K. pneumoniae isolates were dominant in the infection cases compared with those in the colonization/asymptomatic cases, suggesting that rmpA may play a crucial role in the development of urinary tract infection and pneumonia.

Introduction

Klebsiella pneumoniae is the infectious agent in various diseases, such as bacteremia, meningitis, pneumonia, and urinary tract infection (UTI). A hypervirulent variant of K. pneumoniae that causes severe liver abscesses and bacteremia emerged in the 1980s,1–3 which shows the main characteristics of hypermucoviscosity and tendency to cause severe community-acquired infection.4 This variant was initially mainly reported in Asian countries such as Taiwan5 and South Korea,6 but has now spread to countries outside Asia.7–9 This new variant often harbors the K1 or K2 capsular polysaccharide.4

In addition to the K1 or K2 capsular polysaccharide and the hypermucoviscous phenotype, several genes have been reported as severity determinants, including magA and rmpA, based on studies comparing cases with severe infection to those with localized infection, and in murine lethality experiments.10,11 The magA gene, which encodes a polymerase involved in capsule synthesis, first emerged as one of the candidate virulence genes,12 but is now recognized as a K1 surrogate marker.13,14 Moreover, rmpA, which regulates capsular polysaccharide synthesis, was shown to be responsible for virulence in clinical studies10,13,15 and in murine models.16,17 The kfu gene, which mediates the uptake of ferric iron, and allS, encoding the activator of the allantoin regulon,18 were also considered as virulence factors in a murine intraperitoneal challenge model.19,20 Other reported virulence factors of K. pneumoniae include the wabG gene, which plays a role in formation of the outer core lipopolysaccharide, and the uge gene, which encodes a uridine diphosphate galacturonate 4-epimerase and is required for biosynthesis of the capsule and smooth lipopolysaccharide.21 However, the prevalence of these genes did not significantly differ between UTI cases and cases without infection.22 The mrkD gene, which encodes type 3 fimbrial adhesion protein, was reported as a factor related to biofilm formation,23 but its influence on virulence in a clinical setting remains unclear.24

Thus, the factors that cause severe infections, such as bacteremia and liver abscess, have been widely investigated both in clinical studies and in animal models. Moreover, some epidemiological studies of hypervirulent K. pneumoniae strains causing other types of infections, such as pneumonia and UTI, have also been reported.22,25 Furthermore, some of the virulence factors of K. pneumoniae pneumonia associated with bacteremia in Japan were identified.26 Despite this recent focus on the virulence factors underlying severe bacterial infection and their relationship with infection severity, the difference in the prevalence of the hypervirulent K. pneumoniae strains and other strains and the characteristics between infection and colonization/asymptomatic status are still poorly understood.

Therefore, we performed a retrospective study to investigate the prevalence of the hypermucoviscous phenotype and virulent K. pneumoniae strains in a tertiary care teaching hospital in Japan and to determine the differences in infection status with respect to patient background and microbiological characteristics. This study can contribute to gaining a deeper understanding of the epidemiology of virulent K. pneumoniae strains and the virulence factors involved in the development of K. pneumoniae infection.

Patients and methods

Patients and setting

This retrospective study was conducted at the University of Tokyo Hospital, a 1,217-bed tertiary care teaching hospital in Tokyo, Japan. We randomly selected 130 patients among those who tested positive for the presence of K. pneumoniae isolates in their sputum or urine from January 2011 to November 2012. One isolate from the sputum or urine was included from each patient. Patient data, clinical symptoms, and microbiological data were retrieved from the medical records.

Data collection and definitions

Collected patient data included age, sex, underlying disease (diabetes mellitus, malignancy, neutropenia, collagen disease, cirrhosis, chronic kidney disease, and chronic pulmonary disease), and use of immunosuppressant. The presence of intravascular catheters, including peripheral vessel catheters, central venous catheters, and arterial vessel catheters, as well as urinary catheters was reviewed. History of residence in a nursing home and antibiotic use within a month were also reviewed. The samples obtained within 48 hours after admission were defined as community isolates; the other samples were defined as hospital isolates.

The cases were divided into infection and colonization cases. Infection was defined depending on the type of sample that showed the presence of K. pneumoniae; ie, pneumonia in cases from whom sputum was obtained and UTI in cases from whom urine was obtained. Definitions of these infections were compliant with the relevant guidelines.27,28 Pneumonia due to K. pneumoniae was defined as 1) manifestation of clinical symptoms such as fever (>37.5°C) and purulent sputum (assessed by Gram staining or appearance of sputum), 2) compatible radiological findings, and 3) dominant detection of K. pneumoniae in sputum culture. Pneumonia cases for which the samples did not meet all of these criteria were defined as colonization cases. UTI due to K. pneumoniae was defined as 1) manifestation of clinical symptoms such as fever (>37.5°C) and pain during urination, 2) positive leukocyte esterase test in urinary analysis, and 3) single or dominant detection of K. pneumoniae in urine culture. UTI cases for which the samples did not meet all of these criteria were defined as asymptomatic cases.

Microbiological procedures

All isolates were identified using the WalkAway system (Siemens, Berlin, Germany). Antimicrobial susceptibility was determined with the WalkAway system and the standard criteria of the Clinical and Laboratory Standards Institute guide M100-S27.29 The hypermucoviscosity of K. pneumoniae was assessed with the string test, in which a viscous string >5-mm long was considered positive.24,30 The capsular polysaccharide serotype (K1 and K2) was determined by polymerase chain reaction (PCR) for serotype-specific targets within the K1 and K2 cps clusters, as described previously.31 To detect virulence-associated genes, such as magA, allS, rmpA, mrkD, uge, kfu-BC, and wabG, the genomic DNA of K. pneumoniae was extracted and PCR was performed with the specific primers, according to a previous report.32

Statistical analyses

Categorical data were analyzed using the two-tailed Fisher’s exact test and nonparametric data were analyzed using the Mann–Whitney U-test. Differences with p<0.05 were considered significant. All statistical analyses were performed using JMP Pro version 11 software (SAS Institute, Cary, NC, USA).

Ethical considerations

This study was approved by the research ethics committee of the University of Tokyo Hospital. The requirement to obtain written informed consent from each patient was waived because this was an observational retrospective study. The data were analyzed anonymously.

Results

Clinical characteristics of patients

During the study period, 12,467 urine and 4,773 sputum samples were submitted to the clinical laboratory. Among these, 4,093 urine and 1,989 sputum samples were positive for bacteria or fungi. K. pneumoniae was detected in the sputum of 322 cases and in the urine of 288 cases. We randomly enrolled 130 cases (68 in the sputum group and 62 in the urine group) and divided them into infection cases and colonization/asymptomatic cases.

The sputum group comprised 21 pneumonia (infection) and 47 colonization cases. The patients’ characteristics in the infection and colonization groups were similar (Table 1). Two infection cases were complicated by bacteremia. No liver abscess accompanied pneumonia.

Table 1

Characteristics of patients in whom Klebsiella pneumoniae was isolated from the sputum

Variables	Totaln=68	Infection casesn=21	Colonization casesn=47	p-value
Age (mean, range)	69 (1–94)	66 (36–94)	72 (1–92)	0.474
Sex (male/female)	49/19	17/4	32/15	0.384
Hospital/community	41/27	11/10	30/17	0.428
Comorbidity
Diabetes mellitus	19	6	13	1
Malignancy	31	8	23	0.442
Neutropenia	2	1	1	0.526
Collagen disease	6	0	6	0.166
Cirrhosis	2	1	1	0.526
Chronic kidney disease	12	2	10	0.317
Chronic pulmonary disease	16	5	11	1
Immunosuppressant	8	1	7	0.419
Intravascular device	38	12	26	1
Urinary catheter insertion	14	4	10	1
Residence in nursing home	7	3	4	0.668
Antibiotic use within a month	21	5	16	0.571

The urine group comprised 28 UTI (infection) and 34 asymptomatic cases, and the patients’ characteristics were similar between the two groups (Table 2). Bacteremia was present in four cases. No liver abscess was detected in any of the UTI cases.

Table 2

Characteristics of patients in whom Klebsiella pneumoniae was isolated from the urine

Variables	Totaln=62	Infection casesn=28	Asymptomatic casesn=34	p-value
Age (mean, range)	69 (0–91)	74 (0–84)	67 (0–91)	0.579
Sex (male/female)	24/38	9/19	15/19	0.434
Hospital/community	23/39	8/20	15/19	0.292
Comorbidity
Diabetes mellitus	26	11	15	0.798
Malignancy	32	16	16	0.456
Neutropenia	2	1	1	1
Collagen disease	10	6	4	0.326
Cirrhosis	3	0	3	0.245
Chronic kidney disease	11	8	3	0.053
Chronic pulmonary disease	5	3	2	0.650
Immunosuppressant	13	7	6	0.541
Intravascular device	18	5	13	0.097
Urinary catheter insertion	20	10	10	0.785
Residence in nursing home	3	2	1	0.585
Antibiotic use within a month	13	6	7	1

Virulence genes and antimicrobial susceptibility of K. pneumoniae isolates

Of the 130 K. pneumoniae isolates, 25 (19.2%) exhibited capsular serotype K1 or K2 and 33 (25.4%) showed hypermucoviscosity. The prevalence of virulence-associated genes in the sputum and urine groups is shown in Tables 3 and 4. In both the sputum and urine groups, the prevalence of rmpA in the infection cases was significantly higher than that in the colonization/asymptomatic cases (p=0.016 in the sputum group and p=0.037 in the urine group). Prevalence of the other genes did not differ significantly between the infection and colonization/asymptomatic cases in both the sputum and urine groups.

Table 3

Bacterial characteristics of Klebsiella pneumoniae isolated from pneumonia cases

Variables	Totaln=68	Pneumonia casesn=21	Colonization casesn=47	p-value
Serotype K1 or K2	16	6	10	0.546
K1	8	2	6	1
K2	8	4	4	0.240
Hypermucoviscosity	24	10	14	0.178
Virulence factor
magA	7	2	5	1
rmpA	18	10	8	0.016a
allS	8	3	5	0.695
mrkD	52	17	35	0.759
uge	55	20	35	0.051
kfu	18	4	14	0.553
wabG	60	19	41	1

Note:

Statistically significant.

Table 4

Bacterial characteristics of Klebsiella pneumoniae isolated from urinary tract infection cases

Variables	Totaln=62	UTI casesn=28	Asymptomatic casesn=34	p-value
Serotype K1 or K2	9	6	3	0.277
K1	6	5	1	0.082
K2	3	1	2	1
Hypermucoviscosity	9	4	5	1
Virulence factor
magA	6	5	1	0.082
rmpA	4	4	0	0.037a
allS	9	4	5	1
mrkD	59	26	33	0.585
uge	48	21	27	0.765
kfu	30	16	14	0.307
wabG	59	26	33	0.585

Note:

Statistically significant.

Abbreviation: UTI, urinary tract infection.

Of the 130 isolates, antimicrobial susceptibility data for one strain were missing. Except for this isolate, all of the isolates showed 100% sensitivity to ceftazidime, cefozopran, meropenem, gentamicin, and levofloxacin. No extended-spectrum beta-lactamase (ESBL)-producing strains were detected.

Discussion

K. pneumoniae is the etiological agent of pneumonia and UTI. Although the pathogenesis and virulence factors of K. pneumoniae have been widely studied, the differences between infectious cases and colonization/asymptomatic cases are not well understood. Therefore, in this study, we compared the clinical and microbiological characteristics between infection cases and colonization cases in a tertiary care teaching hospital in Japan. We determined the prevalence of hypermucoviscosity, specific serotypes, and virulence genes in K. pneumoniae isolated from the sputum and urine obtained from selected patients. We also showed the significant dominance of rmpA in infection cases in both the sputum and urine groups.

Diabetes mellitus has been reported as a risk factor for hypervirulent K. pneumoniae infection.4,33 However, we did not find a difference in the prevalence of diabetes mellitus between the infection and colonization/asymptomatic cases.

The infection and colonization/asymptomatic groups showed a similar prevalence of serotypes and the hypermucoviscous phenotype; 25.4% and 19.2% of all isolates had the hypermucoviscous phenotype and the K1/K2 serotypes, respectively. Although the epidemiology of hypermucoviscous K. pneumoniae in various types of infection in Japan is not clear, ~10% of bacteremic K. pneumoniae were reported to exhibit the hypermucoviscous phenotype34 and 18.3% of K. pneumoniae isolates causing pneumonia belonged to the K1 and K2 serotypes.26 In Taiwan, approximately half of the K. pneumoniae isolates were found to belong to serotype K1/K2 in patients with community-acquired bacterial pneumonia25,35 and in nasopharynx carriers.35 In China, 33% of the isolates from hospitalized patients were reported to be hypermucoviscous.36 Outside Asia, infection with serotype K1/K2 or hypermucoviscous K. pneumoniae appears to be quite rare. In Europe, 1.1% of 1090 isolates belonged to the K1-CC23 serotype in the UK,37 and 5.4% of 878 isolates were hypermucoviscous, whereas 0.3% were rmpA-positive in Spain.38 In the USA, 6.3% of the 64 isolates obtained from blood cultures were found to be rmpA- or magA-positive.39 These reports indicate that the prevalence of specific serotypes and the hypermucoviscous phenotype vary geographically.

Of the virulence genes investigated in the present study, rmpA was significantly dominant in the infection cases. Indeed, the K1/K2 strains, previously known as virulent strains, were shown to be avirulent without rmpA or the hypermucoviscous phenotype in a murine model.40 Among the non-K1/K2 strains, rmpA-positive strains showed high virulence in murine lethality tests.41 Moreover, in Taiwan, 85.7% and 87.0% of non-K1/K2 isolates were found to be rmpA-positive in cases of liver abscess and bacteremia, respectively.25,40 Although some rmpA-positive isolates did not exhibit the hypermucoviscous phenotype, these isolates were more virulent than the rmpA-negative and nonhypermucoviscous isolates.40 Lin et al22 reported that the prevalence of rmpA was significantly higher in the urinary isolates of UTI cases than that in the fecal isolates obtained from healthy people in Taiwan.

The rmpA gene regulates capsule polysaccharide biosynthesis.16 Three types of rmpA genes have been reported to date: chromosomally located rmpA, rmpA on a plasmid, and rmpA2 on a plasmid.24 Although the primers used in this study could only detect plasmid-mediated rmpA and rmpA2, the strains possessing chromosomally located rmpA appear to be much less prevalent than those possessing plasmid-mediated rmpA/A2.24 Hypercapsule synthesis regulated by rmpA contributes to preventing the phagocytosis and opsonophagocytosis of K. pneumoniae by the host immune cells and inhibits complement-mediated lysis and opsonization.42,43 Thus, rmpA regulation is associated with the escape of K. pneumoniae from immune responses,44 indicating a potentially crucial role in the development of K. pneumoniae infection. Although the prevalence of other reported virulence genes, such as magA, allS, mrkD, uge, kfu, and wabG, did not differ significantly between the infection and colonization groups, we cannot exclude the possibility that these genes might also have some effect in the development of K. pneumoniae infection, especially since not all of the infection cases were associated with an rmpA-positive strain.

No antimicrobial multiresistant strain was detected in our study. A previous study24 reported that hypermucoviscous strains are generally susceptible to antimicrobials except for ampicillin; however, ESBL-producing strains and carbapenem-resistant strains are emerging, especially in China.45,46 Although the prevalence rate of ESBL-producing K. pneumoniae is still low in Japan,47 multi-antimicrobial resistance among hypermucoviscous K. pneumoniae strains may become a future threat.

There are several limitations of this study. First, because this was a retrospective study, it is unknown whether the rmpA-positive isolates in the colonization/asymptomatic group eventually caused infections. Second, this was a single-center analysis and the sputum and urine samples were randomly selected; therefore, not all the samples collected during this time period were analyzed. Thus, a multicenter study is warranted, especially considering that the prevalence of the hypervirulent variant differs from region to region.

Conclusion

In conclusion, the prevalence of hypermucoviscous K. pneumoniae was found to be 25.4%, and rmpA was detected at significantly higher rates in the infection cases than in the colonization/asymptomatic cases in both the sputum and urine groups. The prevalence of serotypes and the hypermucoviscous phenotype is highly variable across studies according to the geographical region. Nevertheless, our study suggests that the rmpA gene may play a crucial role in the development of UTI and pneumonia.