Seroepidemiology of Klebsiella pneumoniae in an Australian Tertiary Hospital and Its Implications for Vaccine Development
http://www.100md.com
微生物临床杂志 2006年第1期
CRC for Vaccine Technology, Herston, Queensland 4006, Australia;
Department of Microbiology & Immunology, University of Melbourne, Victoria 3010, Australia
Australian Bacterial Pathogenesis Program, Clayton, Victoria 3800, Australia
Infectious Diseases Unit
Microbiology Department, The Alfred Hospital, Prahran, Victoria 3181, Australia
Health Protection Agency, Colindale, London NW9 5HT, United Kingdom
CSL Ltd., Parkville, Victoria 3052, Australia
The Burnet Institute, Melbourne, Victoria 3004, Australia
ABSTRACT
The aim of this study was to determine the diversity of Klebsiella pneumoniae capsular serotypes in an Australian setting. Consecutive (n = 293) nonrepetitive isolates of K. pneumoniae from a large teaching hospital laboratory were analyzed. The majority of isolates were from urinary specimens (60.8%); the next most common source was sputum (14.3%), followed by blood (14%). Serotyping revealed a wide range of capsule types. K54 (17.1%), K28 (4.1%), and K17 (3.1%) were the most common, and K54 isolates displayed a high degree of clonality, suggesting a common, nosocomial source. In vitro, one K54 isolate was more adherent to urinary catheters and HEp-2 cells than four other tested isolates; it was slightly more resistant to chlorhexidine but was more susceptible to drying than heavily encapsulated strains. This is the first seroprevalence survey of K. pneumoniae to be performed on Australian isolates, and the high level of diversity of serotypes suggests that capsule-based immunoprophylaxis might not be useful for Australia. In addition there are significant differences in the predominance of specific serotypes compared to the results of surveys performed overseas, which has important implications for capsule-based immunoprophylaxis aimed at a global market.
INTRODUCTION
The aims of this study were to determine the distribution of K serotypes of K. pneumoniae among isolates collected in a tertiary hospital and to determine whether the distribution of capsular types was narrow enough that a vaccine based on capsular antigens was a practical option for prevention and therapy and, in addition, to identify whether more infectious "clones" were present and to link the emergence of such clones with a phenotype that might favor their survival.
K. pneumoniae is a gram-negative bacillus of the family Enterobacteriaceae with a world-wide distribution and is an important cause of human disease resulting in significant morbidity and mortality. The bacterium most typically causes infections of the urinary tract and pneumonia and bacteremia and, less often, wound infections and meningitis that may be acquired both nosocomially and in the wider community (26). K. pneumoniae has been described as an independent predictor for mortality in severe community-acquired pneumonia (24).
K. pneumoniae is considered an extracellular pathogen whose virulence is linked with the production of a polysaccharide capsule that provides protection against host defense mechanisms, particularly phagocytosis (8). Immunity against the encapsulated bacterium is largely mediated by antibodies specific for the capsular polysaccharide, an observation that has been exploited to develop prototypic vaccines against the bacterium (13). As with other capsule-based vaccines, e.g., 23-valent pneumococcus vaccine, the efficacy of similar K. pneumoniae vaccines will depend on the distribution of capsule or "K" serotypes (5, 6). The capsular distribution for K. pneumoniae K types is known to differ worldwide (3, 9, 16, 19, 29), but an acceptable explanation for this phenomenon has yet to be found.
MATERIALS AND METHODS
Over a 13-month period (September 2001 and November 2002) all K. pneumoniae clinical isolates that were identified at the microbiological laboratory of the Alfred Hospital were collected. The laboratory processes specimens from the main hospital (300-bed tertiary referral University teaching hospital) and three other hospitals (a geriatric hospital, a district community hospital, and a hospice). Isolates were identified using a GNI+ card (Vitek; bioMerieux, Marcy l'Etoile, France).
Infections with K. pneumoniae were considered to be community acquired when the isolate was grown from a specimen taken within the first two complete days of admission to hospital. Urinary isolates were associated with significant bacteriuria (>105 CFU/ml) and the presence of white cells in the urine, i.e., >10 cells per high-powered field, unless the patients were neutropenic. Sputum cultures were included when associated with >25 neutrophils per high-powered field on microscopy. Inclusion of other sites (blood, wounds) required the individual to have signs or symptoms of disease.
The isolates were nonrepetitive, i.e., only one isolate was included per patient per episode of infection. When K. pneumoniae was grown simultaneously from samples from different sites for an individual, only one isolate was included in the analysis unless the organisms were of clearly different phenotypes (e.g., serotype). A second isolate for an individual was included when at least 30 days elapsed between the episodes of infection. All isolates were stored at –70°C in Luria-Bertani (LB) glycerol (22%) broth.
Adhesion and drying assays. Five strains of K. pneumoniae were investigated for their ability to adhere to HEp-2 cells and urinary catheter plastic and for their survival after drying in the environment: B5055 (capsule type K2 mouse-lethal strain obtained from the Staten Serum Institut, Denmark), B5055nm (a capsule mutant of B5055 generated in our laboratory; data not shown), and three clinical isolates from the survey, including one of the clonal K54 isolates from a urine specimen and a K1 strain and a K2 strain, both isolated from wound infections.
HEp-2 cells and urinary catheter adhesion assays. Sterile circular (1 cm2) glass coverslips were placed into 24-well tissue culture trays and seeded with approximately 1.8 x 105 HEp-2 cells (in 500 μl minimal essential medium plus 8% fetal calf serum, 20 mM HEPES, 2 mM glutamine, 40 mg/ml gentamicin, and nonessential amino acids) and incubated at 37°C in 6% CO2 for 18 h until the cells were confluent. The cells were then washed with sterile phosphate-buffered saline (PBS), and antibiotic-free medium was added. Approximately 5 x 104 CFU of K. pneumoniae cells were then added to each well and incubated at 37°C in 6% CO2 for 3 h. The medium was then removed, and the adherent cells and bacteria were washed with PBS. The coverslips were transferred to a new 24-well tray and washed again with PBS. A 200-μl volume of 0.1% digitonin was added to cover each coverslip and incubated at room temperature for 5 min, when 800 μl of LB broth was added. Bacterial CFU counts were made by serial dilutions on LB agar plates and are expressed as percentages of the CFU originally applied for each strain. A similar method was used for the urinary catheter adhesion assay, except the bacteria were incubated with 0.5-cm lengths of a sterile 18-gauge Foley urinary catheter.
Drying assay. Approximately 2.0 x 106 CFU bacteria in 10 μl PBS were applied to a sterile glass coverslip and dried in a laminar flow hood for 1, 4, or 24 h. The coverslips were then transferred to tissue culture well and incubated in 200 μl of 0.1% digitonin for 5 min at room temperature before 800 μl of LB broth was added, and the coverslips were vigorously washed with the broth. The surviving bacteria were counted by serial dilutions and expressed as a percentage of the CFU originally applied to the glass coverslip.
Mucoviscosity. Single colonies of the five strains were prepared by overnight culture on LB agar plates and tested for their ability to form viscous strings when a microbiological loop was touched onto their surface and slowly raised. A hyperviscous colony was one which enabled a string length of >0.5 cm to be produced (14). Each test was repeated four times for each isolate on different colonies, and measurements were taken by a second observer.
Susceptibility to disinfectants. These five isolates were cultured in the presence of two different chlorhexidine hand-cleansing preparations. Freshly opened containers of 2% chlorhexidine skin cleanser with tartrazine (Microshield2; Johnson and Johnson, North Ryde, Australia) and 1% chlorhexidine preparation with 70% (vol/vol) ethanol (chlorhexidine 1% hand lotion; Orion Laboratories, Welshpool, Australia) were prepared in serial dilutions with LB broth (1:2 to 1:4,096) (with final chlorhexidine concentrations of 0.5% to 0.00049%) in 96-well microtiter trays. Between 5.5 x 105 and 1.2 x 106 CFU of bacteria were added to each well and cultured at 37°C with shaking. Eight wells were used to test each dilution against the five strains for each chlorhexidine preparation. The wells were observed for the presence of turbidity (4).
Serotyping. The isolates were serotyped by counter-current immunoelectrophoresis (CIE) after the method of Palfreyman (25). Briefly, 107 CFU bacteria/ml were prepared in a 4% formaldehyde-saline solution. Approximately 20 μl of the suspension was placed in a well cut into a 10- by 10-cm agar gel supported by a glass plate. Each isolate was tested against 14 pools of antiserum, with three to six individual antisera in each pool (total, 77 antisera). Approximately 20 μl of the pooled antiserum was placed in a similar well 1 cm from the well containing the bacterial suspension. After 200 V and 50 mA were applied for 90 min, a positive reaction was seen, with a line of precipitation in the gel between the two wells. With each positive pool reaction the isolate was tested against the individual sera making up that pool. When no precipitation was seen, the isolate was considered nontypeable.
Antimicrobial susceptibility. Urinary isolates were tested for susceptibility by disk diffusion testing on Isotest agar (Oxoid, Hampshire, United Kingdom) by use of the antibiotics amoxicillin, amoxicillin-clavulanate, cephalexin, trimethoprim, norfloxacin, and gentamicin. When resistance to a cephalosporin was observed, further disk testing (i.e., augmentation of the zone around ceftazidime with the addition of clavulanic acid) was performed to detect the presence of extended-spectrum beta-lactamase (ESBL)-producing strains. Nonurinary isolates were tested using a GNS card (Vitek; bioMerieux, Marcy l'Etoile, France).
Pulsed-field gel electrophoresis (PFGE). Genomic DNA was prepared as outlined by Kaufmann (18). Equal volumes of bacterial suspensions (2.5 Macfarlane) in SE buffer (75 mM NaCl, 25 mM EDTA, pH 7.5) and 2% low gelling agarose were mixed and dispensed into 50-μl molds. The solidified agarose blocks were incubated in lysis buffer (6 mM Tris-HCl, 100 mM EDTA, 1 M NaCl, 0.5% [wt/vol] Brij 58, 0.2% [wt/vol] sodium deoxycholate, 0.5% [wt/vol] N-lauroyl sarcosine and 1 mM MgCl2 [pH 7.5]) containing 0.5 mg/ml lysozyme at 37°C overnight, followed by overnight incubation at 56°C in 0.5 M EDTA-1% (wt/vol) N-lauroyl sarcosine (pH 9.5) containing 60 μg/ml of proteinase K. After lysis the blocks were washed at least four times in TE buffer (10 mM Tris, 10 mM EDTA, pH 7.5) prior to restriction of the DNA with XbaI (Roche, Hertfordshire, United Kingdom) according to manufacturer's instructions. Fragments were separated through 1.2% agarose in 0.5% TBE (44.5 mM Trizma base, 44.5 mM boric acid, 1 mM EDTA) at 12°C on a CHEF Mapper (Bio-Rad Laboratories Ltd., Hemel, Hempstead, United Kingdom) for 30 h at 6 V/cm, with switching times linearly ramped from 5 to 35 s; lambda concatemers were included as molecular size markers. Following staining with ethidium bromide, TIFF images of the banding patterns were analyzed both visually and using BioNumerics software (Applied Maths, Kortrijk, Belgium).
Statistical analysis. Statistical analysis was performed using Excel:mac (Microsoft, Seattle, WA) and SPSS version 12 (Chicago, IL) and Mann-Whitney or Pearson chi-square testing and 1- and 2-proportion analysis for nonparametric data where appropriate.
RESULTS
Epidemiology. Of the 293 isolates, 178 (60.8%) were from urine, 42 (14.3%) from sputum, 41 (14%) from blood, 18 (6.1%) from central venous catheter tips, 12 (4.1%) from wound swabs, and 2 (0.7%) from tissue culture specimens. Urinary tract infections were significantly more common in women (142 infections in women versus 34 in men; P < 0.001) and occurred in an older population compared to other types of infection (mean age, 69.0 versus 58.8 years; P < 0.001). Of the 42 episode of pneumonia caused by K. pneumoniae, 19 were diagnosed for patients in the intensive care unit, making these patients significantly more likely to have pneumonia than any other infection due to K. pneumoniae (P < 0.001). There were no isolates from cerebrospinal fluid in this series. In total 166 isolates (56.7%) were from nosocomial infections.
Serotyping. Antisera of the recognized 77 serotypes (designated K1 through K74 and K80 through K82 inclusive) were used to analyze the isolates. Fifty-nine serotypes were represented in this series. A total of 151 isolates were positive for just one serotype, and 54 had a positive reaction for more than one serotype; in all, there were 67 serotypes or combinations of serotypes, and 34 of these were represented by just 1 isolate. A total of 88 isolates (30%) were nontypeable (Table 1). This observation indicates that a wide range of different K. pneumoniae strains cause infection in a largely urbanized (i.e., Melbourne) catchment whose populace presents to the Alfred Hospital.
K54 serotype. Over 17% of the isolates expressed the serotype K54; the next largest group was K28, which made up 4.1%. Three of the K54 isolates were found to have ESBL activity. A total of 24 of the K54 isolates were randomly selected for analysis by PFGE. Of these, 22 showed a high degree of similarity, having fewer than three bands different from one another, suggesting they were clonal in origin (31). Figure 1 shows the PFGE results for 20 of the K54 isolates tested following digestion of the DNA with XbaI.
The K54 isolates were more likely to be isolated from patients in the main tertiary referral hospital than the non-K54 isolates (88.2% versus 75.2%; P = 0.043); however, these patients were located in 15 different wards, and no particular point source was identifiable. By multivariate analysis K54 isolates were found to be significantly more likely to have acquired disease nosocomially than the non-K54 isolates (38 of 51 versus 128 of 242; odds ratio [OR], 2.32 [95% confidence interval {CI}, 1.10 to 4.92]; P = 0.028). K54 showed a predilection for some types of infection and not others: of the 18 isolates from central venous catheter tips, 10 (55.6%) were K54 and 62.4% of the non-K54 isolates were of urinary origin whereas only 52.1% of the K54 isolates were cultured from urine (though neither result was significant).
Susceptibility testing. A total of 285 isolates were tested against at least five antibiotics by one of the two methods outlined above. The percentages which were susceptible were as follows: for amoxicillin-ampicillin, 15.5%; for amoxicillin-ampicillin with clavulanate, 94.3%; for gentamicin, 93.3%; for trimethoprim, 83%; for norfloxacin, 91%.
ESBL production was detected in 8.2% of the isolates. ESBLs were present in serotypes 5, 11, 14, 16, 21, 25, 26, 30, 54, and 64 and in nontypeable isolates. Three (6.5%) of the 46 K54 isolates tested were found to have ESBLs, a result which was not significantly different from the percentage seen with the non-K54 isolates (8.3%; P = 0.43). By multivariate analysis the ESBLs were significantly more likely to be isolated in the intensive care unit (OR, 5.22 [CI, 1.17 to 23.25]; P = 0.03). Similarly, there was a trend showing that the ESBLs were more likely to be nosocomial in origin, although this did not quite reach statistical significance (OR, 2.61 [CI, 0.67 to 10.21]; P = 0.082). There was no carbapenem resistance found in this survey's isolates, though metallo-beta-lactamase-producing K. pneumoniae bacteria have subsequently been detected in our laboratory; however, the serotypes of these organisms have not been determined (C. Franklin, personal communication).
Adhesion, drying, disinfectant, and mucoviscosity assays. The K54 isolate was found to be significantly more adherent to HEp-2 cells (P < 0.005) and urinary catheter plastic (P < 0.005) than each of the other four strains tested (Fig. 2 and 3). Each of the five strains tested was increasingly vulnerable to drying as time progressed (Fig. 4). However, B5055 was significantly (P < 0.005) more resistant to drying than K54 and K1 strains at all three time points and was significantly (P < 0.005) more resistant to drying than both B5055nm and K2 at 4 and 24 h, as well. All five K. pneumoniae strains were fully susceptible to both disinfectant products at a concentration of 0.0078%, which suggests that these preparations would be probably successful in killing these organisms during hand washing. However, the K54 strain was able to grow in 15 of the 16 wells at a 0.0019% concentration of chlorhexidine whereas the other strains were more susceptible at that concentration; however, this difference was not significant (Table 2).
B5055, K1, and K2 demonstrated a high degree of mucoviscosity and produced viscous string lengths greater than 0.5 cm (B5055 up to 10 cm, K1 up to 3 cm, and K2 up to 8 cm). B5055nm and K54 strains produced no measurable "string" when touched with a metal loop.
DISCUSSION
This novel survey of K. pneumoniae isolates from a variety of hospitals served by a tertiary hospital microbiology laboratory revealed that the bacterium was responsible for at least 293 infections in a 13-month period, with most of the disease manifestations being nosocomial in origin. K. pneumoniae is recognized as a cause of serious morbidity and can increase lengths of hospital stay and, therefore, cost (1, 22, 30). A seroprevalence study of capsule types has not been performed on a large collection of Australian isolates before. The aim of this study was to determine whether the range of capsular serotypes of K. pneumoniae found in a large urban hospital was sufficiently narrow to support the use of immunoprophylaxis (targeting capsular antigens) as a viable therapeutic option.
The isolates were analyzed for capsular serotype by CIE, a test that is only available in a restricted number of reference laboratories worldwide. Of the 205 isolates that gave a positive reaction for capsule in CIE, 67 patterns (single serotypes or combinations of serotypes) were observed. Some of the isolates reacted in combinations of patterns that have been noted in previous studies (19, 25) (Table 1). Isolates which show cross-reactivity with several anticapsular sera have been thought to have greater resistance to antimicrobials (19); however, this was not the case in this series. Despite the use of enriched media to encourage capsule growth, 30% of the isolates were nontypeable and many of these appeared as small, dry colonies and therefore seemed to lack a capsule entirely. There were other isolates that appeared mucoid but did not react with the antiserum panel, and it is possible these are capsule types not included in the 77 standard K types used by most Klebsiella reference laboratories. Clearly, a wide range of serotypes (and nontypeable isolates) exists in this population and argues against the practicality of vaccine- or immunotherapy-based disease prevention where such a course is based on the capsular polysaccharide. There have been no similar surveys prior to this one in Australia, so the range and distribution of serotypes in other parts of the country are unknown.
A similarly wide range of capsular serotypes has been demonstrated in other studies; however, there are differences in the serotypes that appear most frequently in some countries. In Taiwan the most common serotypes isolated were K1, K2, and K57 (16), whereas in Melbourne K54, K28, and K17 were the commonest capsular types observed. The clinical picture was different too: only 14% of the Melbourne isolates were cultured from the blood, whereas 41.2% of Taiwanese K. pneumoniae isolates were from blood specimens. In Taiwan K. pneumoniae urinary tract infection isolates comprised just 19.3% of the total, whereas over 60% of the Australian isolates were from this site (16). The K54 serotype was clearly the commonest single strain found, and it was overwhelmingly clonal, potentially indicating a common source. However, during the period of the survey there was no "outbreak" of K. pneumoniae infection noted and the overall numbers of infections caused by K. pneumoniae were relatively stable in the Alfred Hospital from 1998 to the present (data not shown).
Antimicrobial susceptibility testing was carried out using a Vitek GNS card or disk susceptibility testing methodology adapted from CLSI recommendations. Although 15.5% were found to be "sensitive" to ampicillin-amoxicillin, these agents would not be recommended for therapeutic use, as the result probably indicates poor induction of the beta-lactamase in vitro. The majority of our series were from nosocomial infections; however, relatively few of the isolates produced ESBLs and these were susceptible to most antibiotics tested for. In neighboring Pacific countries the SENTRY study has determined that in some locations, more than 40% of K. pneumoniae isolates are ESBL positive (2). This has led to research into the use of vaccines for the management of disease (especially nosocomial) caused by this bacterium.
Based on European seroprevalence data for bacteremic isolates, a Swiss group developed a 24-valent capsule-based vaccine that was investigated in a clinical trial of 10 patients who were victims of acute trauma. The vaccine elicited a fourfold rise in antibody titer (compared to baseline) in 80% or more of the patients for 21 of the 24 capsular antigens, but its efficacy was unclear (7). This vaccine did not contain a K54-derived antigen. This, and the fact that 30% of the Australian isolates were nontypeable, would make the usefulness of such a vaccine in an urban Australian setting highly questionable.
A K54 isolate was compared to other isolates for phenotypic traits that might render it more transmissible in the hospital setting. The bacterium was avirulent for mice (intravenous 50% infectious dose, >107 CFU), whereas B5055 and the K1 and K2 clinical isolates (the strains which displayed mucoviscosity in the "string" test) were found to be lethal for mice at an inoculum of approximately104 CFU in a bacteremic infection model (data not shown). The B5055nm strain did not display mucoviscosity and, like K54, was avirulent. It has been reported that although the capsule is an important virulence factor, not all encapsulated strains of the same serotype are equally virulent. One group found that that a virulence plasmid (also coding for the siderophore aerobactin) conferred a "large colony, viscid" phenotype to a particular strain of K2. This phenotype, although it "glistened," did not produce excess amounts of capsule but was 1,000-fold more lethal to Swiss mice than the same K2 strain lacking this plasmid (23). This phenomenon has also been described as "hyperviscosity" and has been associated with clinical isolates that cause invasive disease; microscopically, the bacterial colonies produce an exopolysaccharide web that is attached to the capsule (14). Capsule polysaccharides are recognized as virulence determinants and have the ability to protect the bacterium from phagocytosis (26); they also provide protection against desiccation (27). However, capsules may impede other cellular functions such as adhesion. The K54 isolate was found to be more adherent to surfaces, including plastic and human epithelial cells, than other strains but was not more resistant to drying than the others strains tested. It has been noted that a reduction in capsule can increase the adhesive abilities of K. pneumoniae (15); this may explain why K54 was more adherent than B5055. However, K54 was also significantly more adherent to both HEp-2 cells and plastic than B5055nm, a defined, constructed nonencapsulated bacterium from the parent strain B5055, indicating that more than just the presence of a thick capsule prevented B5055 from adhering to surfaces to the same degree as K54. However, these conclusions are cautiously drawn, as only one clonal K54 isolate was tested against these K1 and K2 (and K2 mutant) strains.
As well as having type 1 fimbriae, most K. pneumoniae isolates have type 3 fimbriae, the major subunit being the Mrk A polypeptide that allows adsorption to abiotic polymers of medical devices. Some strains also possess the Mrk D adhesin, which enables the bacteria to adhere and replicate on these polymers (11, 17, 28). There are several reports linking adhesive phenotypes of K. pneumoniae to nosocomial infection (10, 12, 20). Adhesion to fomites is an advantage for survival in the hospital environment and establishment of infection in medical devices, and this may have been a factor in the high proportion of central venous catheter infections found to be due to K54. In addition, adhesion to epithelial cells might allow for persistent colonization of the gastrointestinal tract, supporting resistance to the powerful waves of peristalsis (21). The K54 isolate was slightly more resistant to common hospital disinfectants, though whether this helps to explain the prevalence of this serotype has yet to be determined.
K. pneumoniae is an important cause of morbidity worldwide and shows a wide range of capsular serotypes in surveys from many countries, including this Australian study. Studies should continue to determine why some isolates of bacteria such as K. pneumoniae persist in causing disease in hospital environments. The increase in ESBL-producing and carbapenem-resistant K. pneumoniae isolates means that attempts should be made to define noncapsular targets which might form the basis of vaccines or immunoprophylaxis to protect frequently hospitalized patients against this bacterium. If a vaccine were to be developed, it would be infeasible to base it on capsular antigens, as the targets required are too numerous.
ACKNOWLEDGMENTS
We are grateful for the assistance of Anne Drake, Chris Lemoh, Ben Howden, Rick Corsi, Leanne Arnold, Phuong Pham, and particularly Clare Franklin (The Alfred Hospital), Martin Pearse and Sterling Edwards (CSL Ltd./CRC-VT), Jane Turton and Lyn Whittington (HPA), and Jason Price, Dianna Hocking, Te-Chieh Hung, and Marnie Collins (University of Melbourne). CSL Ltd. is the industry partner of the CRC for Vaccine Technology.
REFERENCES
Anonymous. 2002. The cost of antibiotic resistance: effect of resistance among Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa on length of hospital stay. Infect. Control Hosp. Epidemiol. 23:106-108.
Bell, J. M., J. D. Turnidge, A. C. Gales, M. A. Pfaller, and R. N. Jones. 2002. Prevalence of extended spectrum beta-lactamase (ESBL)-producing clinical isolates in the Asia-Pacific region and South Africa: regional results from SENTRY Antimicrobial Surveillance Program (1998-1999). Diagn. Microbiol. Infect. Dis. 42:193-198.
Blanchette, E. A., and S. J. Rubin. 1980. Seroepidemiology of clinical isolates of Klebsiella in Connecticut. J. Clin. Microbiol. 11:474-478.
Brooks, S. E., M. A. Walczak, S. Malcolm, and R. Hameed. 2004. Intrinsic Klebsiella pneumoniae contamination of liquid germicidal hand soap containing chlorhexidine. Infect. Control Hosp. Epidemiol. 25:883-885.
Butler, J. C., R. F. Breiman, H. B. Lipman, J. Hofmann, and R. R. Facklam. 1995. Serotype distribution of Streptococcus pneumoniae infections among preschool children in the United States, 1978-1994: implications for development of a conjugate vaccine. J. Infect. Dis. 171:885-889.
Butler, J. C., J. Hofmann, M. S. Cetron, J. A. Elliott, R. R. Facklam, and R. F. Breiman. 1996. The continued emergence of drug-resistant Streptococcus pneumoniae in the United States: an update from the Centers for Disease Control and Prevention's Pneumococcal Sentinel Surveillance System. J. Infect. Dis. 174:986-993.
Campbell, W. N., E. Hendrix, S. Cryz, Jr., and A. S. Cross. 1996. Immunogenicity of a 24-valent Klebsiella capsular polysaccharide vaccine and an eight-valent Pseudomonas O-polysaccharide conjugate vaccine administered to victims of acute trauma. Clin. Infect. Dis. 23:179-181.
Cryz, S. J., Jr., F. Furer, and R. Germanier. 1984. Experimental Klebsiella pneumoniae burn wound sepsis: role of capsular polysaccharide. Infect. Immun. 43:440-441.
Cryz, S. J., Jr., P. M. Mortimer, V. Mansfield, and R. Germanier. 1986. Seroepidemiology of Klebsiella bacteremic isolates and implications for vaccine development. J. Clin. Microbiol. 23:687-690.
Di Martino, P., Y. Bertin, J. P. Girardeau, V. Livrelli, B. Joly, and A. Darfeuille-Michaud. 1995. Molecular characterization and adhesive properties of CF29K, an adhesin of Klebsiella pneumoniae strains involved in nosocomial infections. Infect. Immun. 63:4336-4344.
Di Martino, P., N. Cafferini, B. Joly, and A. Darfeuille-Michaud. 2003. Klebsiella pneumoniae type 3 pili facilitate adherence and biofilm formation on abiotic surfaces. Res. Microbiol. 154:9-16.
Di Martino, P., V. Livrelli, D. Sirot, B. Joly, and A. Darfeuille-Michaud. 1996. A new fimbrial antigen harbored by CAZ-5/SHV-4-producing Klebsiella pneumoniae strains involved in nosocomial infections. Infect. Immun. 64:2266-2273.
Edelman, R., D. N. Taylor, S. S. Wasserman, J. B. McClain, A. S. Cross, J. C. Sadoff, J. U. Que, and S. J. Cryz. 1994. Phase 1 trial of a 24-valent Klebsiella capsular polysaccharide vaccine and an eight-valent Pseudomonas O-polysaccharide conjugate vaccine administered simultaneously. Vaccine 12:1288-1294.
Fang, C. T., Y. P. Chuang, C. T. Shun, S. C. Chang, and J. T. Wang. 2004. A novel virulence gene in Klebsiella pneumoniae strains causing primary liver abscess and septic metastatic complications. J. Exp. Med. 199:697-705.
Favre-Bonte, S., B. Joly, and C. Forestier. 1999. Consequences of reduction of Klebsiella pneumoniae capsule expression on interactions of this bacterium with epithelial cells. Infect. Immun. 67:554-561.
Fung, C. P., B. S. Hu, F. Y. Chang, S. C. Lee, B. I. Kuo, M. Ho, L. K. Siu, and C. Y. Liu. 2000. A 5-year study of the seroepidemiology of Klebsiella pneumoniae: high prevalence of capsular serotype K1 in Taiwan and implication for vaccine efficacy. J. Infect. Dis. 181:2075-2079.
Jagnow, J., and S. Clegg. 2003. Klebsiella pneumoniae MrkD-mediated biofilm formation on extracellular matrix- and collagen-coated surfaces. Microbiology 149:2397-2405.
Kaufmann, M. 1998. Pulsed-field gel electrophoresis, p. 33-50. In N. Woodford and A. Johnson (ed.), Methods in molecular medicine 15: molecular bacteriology: protocols and clinical applications. Humana Press, Totowa, N.J.
Legakis, N. J., L. S. Tzouvelekis, G. Hatzoudis, E. Tzelepi, A. Gourkou, T. L. Pitt, and A. C. Vatopoulos. 1995. Klebsiella pneumoniae infections in Greek hospitals. Dissemination of plasmids encoding an SHV-5 type beta-lactamase. J. Hosp. Infect. 31:177-187.
Livrelli, V., C. De Champs, P. Di Martino, A. Darfeuille-Michaud, C. Forestier, and B. Joly. 1996. Adhesive properties and antibiotic resistance of Klebsiella, Enterobacter, and Serratia clinical isolates involved in nosocomial infections. J. Clin. Microbiol. 34:1963-1969.
Maroncle, N., D. Balestrino, C. Rich, and C. Forestier. 2002. Identification of Klebsiella pneumoniae genes involved in intestinal colonization and adhesion using signature-tagged mutagenesis. Infect. Immun. 70:4729-4734.
Merchant, M., D. R. Karnad, and A. A. Kanbur. 1998. Incidence of nosocomial pneumonia in a medical intensive care unit and general medical ward patients in a public hospital in Bombay, India. J. Hosp. Infect. 39:143-148.
Nassif, X., J. M. Fournier, J. Arondel, and P. J. Sansonetti. 1989. Mucoid phenotype of Klebsiella pneumoniae is a plasmid-encoded virulence factor. Infect. Immun. 57:546-552.
Paganin, F., F. Lilienthal, A. Bourdin, N. Lugagne, F. Tixier, R. Genin, and J. L. Yvin. 2004. Severe community-acquired pneumonia: assessment of microbial aetiology as mortality factor. Eur. Respir. J. 24:779-785.
Palfreyman, J. M. 1978. Klebsiella serotyping by counter-current immunoelectrophoresis. J. Hyg. (London) 81:219-225.
Podschun, R., and U. Ullmann. 1998. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin. Microbiol. Rev. 11:589-603.
Schembri, M. A., D. Dalsgaard, and P. Klemm. 2004. Capsule shields the function of short bacterial adhesins. J. Bacteriol. 186:1249-1257.
Schurtz, T. A., D. B. Hornick, T. K. Korhonen, and S. Clegg. 1994. The type 3 fimbrial adhesin gene (mrkD) of Klebsiella species is not conserved among all fimbriate strains. Infect. Immun. 62:4186-4191.
Smith, S. M., J. T. Digori, and R. H. Eng. 1982. Epidemiology of Klebsiella antibiotic resistance and serotypes. J. Clin. Microbiol. 16:868-873.
Tambyah, P. A., V. Knasinski, and D. G. Maki. 2002. The direct costs of nosocomial catheter-associated urinary tract infection in the era of managed care. Infect. Control Hosp. Epidemiol. 23:27-31.
Tenover, F. C., R. D. Arbeit, R. V. Goering, P. A. Mickelsen, B. E. Murray, D. H. Persing, and B. Swaminathan. 1995. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J. Clin. Microbiol. 33:2233-2239.(Adam W. Jenney, Abigail C)
Department of Microbiology & Immunology, University of Melbourne, Victoria 3010, Australia
Australian Bacterial Pathogenesis Program, Clayton, Victoria 3800, Australia
Infectious Diseases Unit
Microbiology Department, The Alfred Hospital, Prahran, Victoria 3181, Australia
Health Protection Agency, Colindale, London NW9 5HT, United Kingdom
CSL Ltd., Parkville, Victoria 3052, Australia
The Burnet Institute, Melbourne, Victoria 3004, Australia
ABSTRACT
The aim of this study was to determine the diversity of Klebsiella pneumoniae capsular serotypes in an Australian setting. Consecutive (n = 293) nonrepetitive isolates of K. pneumoniae from a large teaching hospital laboratory were analyzed. The majority of isolates were from urinary specimens (60.8%); the next most common source was sputum (14.3%), followed by blood (14%). Serotyping revealed a wide range of capsule types. K54 (17.1%), K28 (4.1%), and K17 (3.1%) were the most common, and K54 isolates displayed a high degree of clonality, suggesting a common, nosocomial source. In vitro, one K54 isolate was more adherent to urinary catheters and HEp-2 cells than four other tested isolates; it was slightly more resistant to chlorhexidine but was more susceptible to drying than heavily encapsulated strains. This is the first seroprevalence survey of K. pneumoniae to be performed on Australian isolates, and the high level of diversity of serotypes suggests that capsule-based immunoprophylaxis might not be useful for Australia. In addition there are significant differences in the predominance of specific serotypes compared to the results of surveys performed overseas, which has important implications for capsule-based immunoprophylaxis aimed at a global market.
INTRODUCTION
The aims of this study were to determine the distribution of K serotypes of K. pneumoniae among isolates collected in a tertiary hospital and to determine whether the distribution of capsular types was narrow enough that a vaccine based on capsular antigens was a practical option for prevention and therapy and, in addition, to identify whether more infectious "clones" were present and to link the emergence of such clones with a phenotype that might favor their survival.
K. pneumoniae is a gram-negative bacillus of the family Enterobacteriaceae with a world-wide distribution and is an important cause of human disease resulting in significant morbidity and mortality. The bacterium most typically causes infections of the urinary tract and pneumonia and bacteremia and, less often, wound infections and meningitis that may be acquired both nosocomially and in the wider community (26). K. pneumoniae has been described as an independent predictor for mortality in severe community-acquired pneumonia (24).
K. pneumoniae is considered an extracellular pathogen whose virulence is linked with the production of a polysaccharide capsule that provides protection against host defense mechanisms, particularly phagocytosis (8). Immunity against the encapsulated bacterium is largely mediated by antibodies specific for the capsular polysaccharide, an observation that has been exploited to develop prototypic vaccines against the bacterium (13). As with other capsule-based vaccines, e.g., 23-valent pneumococcus vaccine, the efficacy of similar K. pneumoniae vaccines will depend on the distribution of capsule or "K" serotypes (5, 6). The capsular distribution for K. pneumoniae K types is known to differ worldwide (3, 9, 16, 19, 29), but an acceptable explanation for this phenomenon has yet to be found.
MATERIALS AND METHODS
Over a 13-month period (September 2001 and November 2002) all K. pneumoniae clinical isolates that were identified at the microbiological laboratory of the Alfred Hospital were collected. The laboratory processes specimens from the main hospital (300-bed tertiary referral University teaching hospital) and three other hospitals (a geriatric hospital, a district community hospital, and a hospice). Isolates were identified using a GNI+ card (Vitek; bioMerieux, Marcy l'Etoile, France).
Infections with K. pneumoniae were considered to be community acquired when the isolate was grown from a specimen taken within the first two complete days of admission to hospital. Urinary isolates were associated with significant bacteriuria (>105 CFU/ml) and the presence of white cells in the urine, i.e., >10 cells per high-powered field, unless the patients were neutropenic. Sputum cultures were included when associated with >25 neutrophils per high-powered field on microscopy. Inclusion of other sites (blood, wounds) required the individual to have signs or symptoms of disease.
The isolates were nonrepetitive, i.e., only one isolate was included per patient per episode of infection. When K. pneumoniae was grown simultaneously from samples from different sites for an individual, only one isolate was included in the analysis unless the organisms were of clearly different phenotypes (e.g., serotype). A second isolate for an individual was included when at least 30 days elapsed between the episodes of infection. All isolates were stored at –70°C in Luria-Bertani (LB) glycerol (22%) broth.
Adhesion and drying assays. Five strains of K. pneumoniae were investigated for their ability to adhere to HEp-2 cells and urinary catheter plastic and for their survival after drying in the environment: B5055 (capsule type K2 mouse-lethal strain obtained from the Staten Serum Institut, Denmark), B5055nm (a capsule mutant of B5055 generated in our laboratory; data not shown), and three clinical isolates from the survey, including one of the clonal K54 isolates from a urine specimen and a K1 strain and a K2 strain, both isolated from wound infections.
HEp-2 cells and urinary catheter adhesion assays. Sterile circular (1 cm2) glass coverslips were placed into 24-well tissue culture trays and seeded with approximately 1.8 x 105 HEp-2 cells (in 500 μl minimal essential medium plus 8% fetal calf serum, 20 mM HEPES, 2 mM glutamine, 40 mg/ml gentamicin, and nonessential amino acids) and incubated at 37°C in 6% CO2 for 18 h until the cells were confluent. The cells were then washed with sterile phosphate-buffered saline (PBS), and antibiotic-free medium was added. Approximately 5 x 104 CFU of K. pneumoniae cells were then added to each well and incubated at 37°C in 6% CO2 for 3 h. The medium was then removed, and the adherent cells and bacteria were washed with PBS. The coverslips were transferred to a new 24-well tray and washed again with PBS. A 200-μl volume of 0.1% digitonin was added to cover each coverslip and incubated at room temperature for 5 min, when 800 μl of LB broth was added. Bacterial CFU counts were made by serial dilutions on LB agar plates and are expressed as percentages of the CFU originally applied for each strain. A similar method was used for the urinary catheter adhesion assay, except the bacteria were incubated with 0.5-cm lengths of a sterile 18-gauge Foley urinary catheter.
Drying assay. Approximately 2.0 x 106 CFU bacteria in 10 μl PBS were applied to a sterile glass coverslip and dried in a laminar flow hood for 1, 4, or 24 h. The coverslips were then transferred to tissue culture well and incubated in 200 μl of 0.1% digitonin for 5 min at room temperature before 800 μl of LB broth was added, and the coverslips were vigorously washed with the broth. The surviving bacteria were counted by serial dilutions and expressed as a percentage of the CFU originally applied to the glass coverslip.
Mucoviscosity. Single colonies of the five strains were prepared by overnight culture on LB agar plates and tested for their ability to form viscous strings when a microbiological loop was touched onto their surface and slowly raised. A hyperviscous colony was one which enabled a string length of >0.5 cm to be produced (14). Each test was repeated four times for each isolate on different colonies, and measurements were taken by a second observer.
Susceptibility to disinfectants. These five isolates were cultured in the presence of two different chlorhexidine hand-cleansing preparations. Freshly opened containers of 2% chlorhexidine skin cleanser with tartrazine (Microshield2; Johnson and Johnson, North Ryde, Australia) and 1% chlorhexidine preparation with 70% (vol/vol) ethanol (chlorhexidine 1% hand lotion; Orion Laboratories, Welshpool, Australia) were prepared in serial dilutions with LB broth (1:2 to 1:4,096) (with final chlorhexidine concentrations of 0.5% to 0.00049%) in 96-well microtiter trays. Between 5.5 x 105 and 1.2 x 106 CFU of bacteria were added to each well and cultured at 37°C with shaking. Eight wells were used to test each dilution against the five strains for each chlorhexidine preparation. The wells were observed for the presence of turbidity (4).
Serotyping. The isolates were serotyped by counter-current immunoelectrophoresis (CIE) after the method of Palfreyman (25). Briefly, 107 CFU bacteria/ml were prepared in a 4% formaldehyde-saline solution. Approximately 20 μl of the suspension was placed in a well cut into a 10- by 10-cm agar gel supported by a glass plate. Each isolate was tested against 14 pools of antiserum, with three to six individual antisera in each pool (total, 77 antisera). Approximately 20 μl of the pooled antiserum was placed in a similar well 1 cm from the well containing the bacterial suspension. After 200 V and 50 mA were applied for 90 min, a positive reaction was seen, with a line of precipitation in the gel between the two wells. With each positive pool reaction the isolate was tested against the individual sera making up that pool. When no precipitation was seen, the isolate was considered nontypeable.
Antimicrobial susceptibility. Urinary isolates were tested for susceptibility by disk diffusion testing on Isotest agar (Oxoid, Hampshire, United Kingdom) by use of the antibiotics amoxicillin, amoxicillin-clavulanate, cephalexin, trimethoprim, norfloxacin, and gentamicin. When resistance to a cephalosporin was observed, further disk testing (i.e., augmentation of the zone around ceftazidime with the addition of clavulanic acid) was performed to detect the presence of extended-spectrum beta-lactamase (ESBL)-producing strains. Nonurinary isolates were tested using a GNS card (Vitek; bioMerieux, Marcy l'Etoile, France).
Pulsed-field gel electrophoresis (PFGE). Genomic DNA was prepared as outlined by Kaufmann (18). Equal volumes of bacterial suspensions (2.5 Macfarlane) in SE buffer (75 mM NaCl, 25 mM EDTA, pH 7.5) and 2% low gelling agarose were mixed and dispensed into 50-μl molds. The solidified agarose blocks were incubated in lysis buffer (6 mM Tris-HCl, 100 mM EDTA, 1 M NaCl, 0.5% [wt/vol] Brij 58, 0.2% [wt/vol] sodium deoxycholate, 0.5% [wt/vol] N-lauroyl sarcosine and 1 mM MgCl2 [pH 7.5]) containing 0.5 mg/ml lysozyme at 37°C overnight, followed by overnight incubation at 56°C in 0.5 M EDTA-1% (wt/vol) N-lauroyl sarcosine (pH 9.5) containing 60 μg/ml of proteinase K. After lysis the blocks were washed at least four times in TE buffer (10 mM Tris, 10 mM EDTA, pH 7.5) prior to restriction of the DNA with XbaI (Roche, Hertfordshire, United Kingdom) according to manufacturer's instructions. Fragments were separated through 1.2% agarose in 0.5% TBE (44.5 mM Trizma base, 44.5 mM boric acid, 1 mM EDTA) at 12°C on a CHEF Mapper (Bio-Rad Laboratories Ltd., Hemel, Hempstead, United Kingdom) for 30 h at 6 V/cm, with switching times linearly ramped from 5 to 35 s; lambda concatemers were included as molecular size markers. Following staining with ethidium bromide, TIFF images of the banding patterns were analyzed both visually and using BioNumerics software (Applied Maths, Kortrijk, Belgium).
Statistical analysis. Statistical analysis was performed using Excel:mac (Microsoft, Seattle, WA) and SPSS version 12 (Chicago, IL) and Mann-Whitney or Pearson chi-square testing and 1- and 2-proportion analysis for nonparametric data where appropriate.
RESULTS
Epidemiology. Of the 293 isolates, 178 (60.8%) were from urine, 42 (14.3%) from sputum, 41 (14%) from blood, 18 (6.1%) from central venous catheter tips, 12 (4.1%) from wound swabs, and 2 (0.7%) from tissue culture specimens. Urinary tract infections were significantly more common in women (142 infections in women versus 34 in men; P < 0.001) and occurred in an older population compared to other types of infection (mean age, 69.0 versus 58.8 years; P < 0.001). Of the 42 episode of pneumonia caused by K. pneumoniae, 19 were diagnosed for patients in the intensive care unit, making these patients significantly more likely to have pneumonia than any other infection due to K. pneumoniae (P < 0.001). There were no isolates from cerebrospinal fluid in this series. In total 166 isolates (56.7%) were from nosocomial infections.
Serotyping. Antisera of the recognized 77 serotypes (designated K1 through K74 and K80 through K82 inclusive) were used to analyze the isolates. Fifty-nine serotypes were represented in this series. A total of 151 isolates were positive for just one serotype, and 54 had a positive reaction for more than one serotype; in all, there were 67 serotypes or combinations of serotypes, and 34 of these were represented by just 1 isolate. A total of 88 isolates (30%) were nontypeable (Table 1). This observation indicates that a wide range of different K. pneumoniae strains cause infection in a largely urbanized (i.e., Melbourne) catchment whose populace presents to the Alfred Hospital.
K54 serotype. Over 17% of the isolates expressed the serotype K54; the next largest group was K28, which made up 4.1%. Three of the K54 isolates were found to have ESBL activity. A total of 24 of the K54 isolates were randomly selected for analysis by PFGE. Of these, 22 showed a high degree of similarity, having fewer than three bands different from one another, suggesting they were clonal in origin (31). Figure 1 shows the PFGE results for 20 of the K54 isolates tested following digestion of the DNA with XbaI.
The K54 isolates were more likely to be isolated from patients in the main tertiary referral hospital than the non-K54 isolates (88.2% versus 75.2%; P = 0.043); however, these patients were located in 15 different wards, and no particular point source was identifiable. By multivariate analysis K54 isolates were found to be significantly more likely to have acquired disease nosocomially than the non-K54 isolates (38 of 51 versus 128 of 242; odds ratio [OR], 2.32 [95% confidence interval {CI}, 1.10 to 4.92]; P = 0.028). K54 showed a predilection for some types of infection and not others: of the 18 isolates from central venous catheter tips, 10 (55.6%) were K54 and 62.4% of the non-K54 isolates were of urinary origin whereas only 52.1% of the K54 isolates were cultured from urine (though neither result was significant).
Susceptibility testing. A total of 285 isolates were tested against at least five antibiotics by one of the two methods outlined above. The percentages which were susceptible were as follows: for amoxicillin-ampicillin, 15.5%; for amoxicillin-ampicillin with clavulanate, 94.3%; for gentamicin, 93.3%; for trimethoprim, 83%; for norfloxacin, 91%.
ESBL production was detected in 8.2% of the isolates. ESBLs were present in serotypes 5, 11, 14, 16, 21, 25, 26, 30, 54, and 64 and in nontypeable isolates. Three (6.5%) of the 46 K54 isolates tested were found to have ESBLs, a result which was not significantly different from the percentage seen with the non-K54 isolates (8.3%; P = 0.43). By multivariate analysis the ESBLs were significantly more likely to be isolated in the intensive care unit (OR, 5.22 [CI, 1.17 to 23.25]; P = 0.03). Similarly, there was a trend showing that the ESBLs were more likely to be nosocomial in origin, although this did not quite reach statistical significance (OR, 2.61 [CI, 0.67 to 10.21]; P = 0.082). There was no carbapenem resistance found in this survey's isolates, though metallo-beta-lactamase-producing K. pneumoniae bacteria have subsequently been detected in our laboratory; however, the serotypes of these organisms have not been determined (C. Franklin, personal communication).
Adhesion, drying, disinfectant, and mucoviscosity assays. The K54 isolate was found to be significantly more adherent to HEp-2 cells (P < 0.005) and urinary catheter plastic (P < 0.005) than each of the other four strains tested (Fig. 2 and 3). Each of the five strains tested was increasingly vulnerable to drying as time progressed (Fig. 4). However, B5055 was significantly (P < 0.005) more resistant to drying than K54 and K1 strains at all three time points and was significantly (P < 0.005) more resistant to drying than both B5055nm and K2 at 4 and 24 h, as well. All five K. pneumoniae strains were fully susceptible to both disinfectant products at a concentration of 0.0078%, which suggests that these preparations would be probably successful in killing these organisms during hand washing. However, the K54 strain was able to grow in 15 of the 16 wells at a 0.0019% concentration of chlorhexidine whereas the other strains were more susceptible at that concentration; however, this difference was not significant (Table 2).
B5055, K1, and K2 demonstrated a high degree of mucoviscosity and produced viscous string lengths greater than 0.5 cm (B5055 up to 10 cm, K1 up to 3 cm, and K2 up to 8 cm). B5055nm and K54 strains produced no measurable "string" when touched with a metal loop.
DISCUSSION
This novel survey of K. pneumoniae isolates from a variety of hospitals served by a tertiary hospital microbiology laboratory revealed that the bacterium was responsible for at least 293 infections in a 13-month period, with most of the disease manifestations being nosocomial in origin. K. pneumoniae is recognized as a cause of serious morbidity and can increase lengths of hospital stay and, therefore, cost (1, 22, 30). A seroprevalence study of capsule types has not been performed on a large collection of Australian isolates before. The aim of this study was to determine whether the range of capsular serotypes of K. pneumoniae found in a large urban hospital was sufficiently narrow to support the use of immunoprophylaxis (targeting capsular antigens) as a viable therapeutic option.
The isolates were analyzed for capsular serotype by CIE, a test that is only available in a restricted number of reference laboratories worldwide. Of the 205 isolates that gave a positive reaction for capsule in CIE, 67 patterns (single serotypes or combinations of serotypes) were observed. Some of the isolates reacted in combinations of patterns that have been noted in previous studies (19, 25) (Table 1). Isolates which show cross-reactivity with several anticapsular sera have been thought to have greater resistance to antimicrobials (19); however, this was not the case in this series. Despite the use of enriched media to encourage capsule growth, 30% of the isolates were nontypeable and many of these appeared as small, dry colonies and therefore seemed to lack a capsule entirely. There were other isolates that appeared mucoid but did not react with the antiserum panel, and it is possible these are capsule types not included in the 77 standard K types used by most Klebsiella reference laboratories. Clearly, a wide range of serotypes (and nontypeable isolates) exists in this population and argues against the practicality of vaccine- or immunotherapy-based disease prevention where such a course is based on the capsular polysaccharide. There have been no similar surveys prior to this one in Australia, so the range and distribution of serotypes in other parts of the country are unknown.
A similarly wide range of capsular serotypes has been demonstrated in other studies; however, there are differences in the serotypes that appear most frequently in some countries. In Taiwan the most common serotypes isolated were K1, K2, and K57 (16), whereas in Melbourne K54, K28, and K17 were the commonest capsular types observed. The clinical picture was different too: only 14% of the Melbourne isolates were cultured from the blood, whereas 41.2% of Taiwanese K. pneumoniae isolates were from blood specimens. In Taiwan K. pneumoniae urinary tract infection isolates comprised just 19.3% of the total, whereas over 60% of the Australian isolates were from this site (16). The K54 serotype was clearly the commonest single strain found, and it was overwhelmingly clonal, potentially indicating a common source. However, during the period of the survey there was no "outbreak" of K. pneumoniae infection noted and the overall numbers of infections caused by K. pneumoniae were relatively stable in the Alfred Hospital from 1998 to the present (data not shown).
Antimicrobial susceptibility testing was carried out using a Vitek GNS card or disk susceptibility testing methodology adapted from CLSI recommendations. Although 15.5% were found to be "sensitive" to ampicillin-amoxicillin, these agents would not be recommended for therapeutic use, as the result probably indicates poor induction of the beta-lactamase in vitro. The majority of our series were from nosocomial infections; however, relatively few of the isolates produced ESBLs and these were susceptible to most antibiotics tested for. In neighboring Pacific countries the SENTRY study has determined that in some locations, more than 40% of K. pneumoniae isolates are ESBL positive (2). This has led to research into the use of vaccines for the management of disease (especially nosocomial) caused by this bacterium.
Based on European seroprevalence data for bacteremic isolates, a Swiss group developed a 24-valent capsule-based vaccine that was investigated in a clinical trial of 10 patients who were victims of acute trauma. The vaccine elicited a fourfold rise in antibody titer (compared to baseline) in 80% or more of the patients for 21 of the 24 capsular antigens, but its efficacy was unclear (7). This vaccine did not contain a K54-derived antigen. This, and the fact that 30% of the Australian isolates were nontypeable, would make the usefulness of such a vaccine in an urban Australian setting highly questionable.
A K54 isolate was compared to other isolates for phenotypic traits that might render it more transmissible in the hospital setting. The bacterium was avirulent for mice (intravenous 50% infectious dose, >107 CFU), whereas B5055 and the K1 and K2 clinical isolates (the strains which displayed mucoviscosity in the "string" test) were found to be lethal for mice at an inoculum of approximately104 CFU in a bacteremic infection model (data not shown). The B5055nm strain did not display mucoviscosity and, like K54, was avirulent. It has been reported that although the capsule is an important virulence factor, not all encapsulated strains of the same serotype are equally virulent. One group found that that a virulence plasmid (also coding for the siderophore aerobactin) conferred a "large colony, viscid" phenotype to a particular strain of K2. This phenotype, although it "glistened," did not produce excess amounts of capsule but was 1,000-fold more lethal to Swiss mice than the same K2 strain lacking this plasmid (23). This phenomenon has also been described as "hyperviscosity" and has been associated with clinical isolates that cause invasive disease; microscopically, the bacterial colonies produce an exopolysaccharide web that is attached to the capsule (14). Capsule polysaccharides are recognized as virulence determinants and have the ability to protect the bacterium from phagocytosis (26); they also provide protection against desiccation (27). However, capsules may impede other cellular functions such as adhesion. The K54 isolate was found to be more adherent to surfaces, including plastic and human epithelial cells, than other strains but was not more resistant to drying than the others strains tested. It has been noted that a reduction in capsule can increase the adhesive abilities of K. pneumoniae (15); this may explain why K54 was more adherent than B5055. However, K54 was also significantly more adherent to both HEp-2 cells and plastic than B5055nm, a defined, constructed nonencapsulated bacterium from the parent strain B5055, indicating that more than just the presence of a thick capsule prevented B5055 from adhering to surfaces to the same degree as K54. However, these conclusions are cautiously drawn, as only one clonal K54 isolate was tested against these K1 and K2 (and K2 mutant) strains.
As well as having type 1 fimbriae, most K. pneumoniae isolates have type 3 fimbriae, the major subunit being the Mrk A polypeptide that allows adsorption to abiotic polymers of medical devices. Some strains also possess the Mrk D adhesin, which enables the bacteria to adhere and replicate on these polymers (11, 17, 28). There are several reports linking adhesive phenotypes of K. pneumoniae to nosocomial infection (10, 12, 20). Adhesion to fomites is an advantage for survival in the hospital environment and establishment of infection in medical devices, and this may have been a factor in the high proportion of central venous catheter infections found to be due to K54. In addition, adhesion to epithelial cells might allow for persistent colonization of the gastrointestinal tract, supporting resistance to the powerful waves of peristalsis (21). The K54 isolate was slightly more resistant to common hospital disinfectants, though whether this helps to explain the prevalence of this serotype has yet to be determined.
K. pneumoniae is an important cause of morbidity worldwide and shows a wide range of capsular serotypes in surveys from many countries, including this Australian study. Studies should continue to determine why some isolates of bacteria such as K. pneumoniae persist in causing disease in hospital environments. The increase in ESBL-producing and carbapenem-resistant K. pneumoniae isolates means that attempts should be made to define noncapsular targets which might form the basis of vaccines or immunoprophylaxis to protect frequently hospitalized patients against this bacterium. If a vaccine were to be developed, it would be infeasible to base it on capsular antigens, as the targets required are too numerous.
ACKNOWLEDGMENTS
We are grateful for the assistance of Anne Drake, Chris Lemoh, Ben Howden, Rick Corsi, Leanne Arnold, Phuong Pham, and particularly Clare Franklin (The Alfred Hospital), Martin Pearse and Sterling Edwards (CSL Ltd./CRC-VT), Jane Turton and Lyn Whittington (HPA), and Jason Price, Dianna Hocking, Te-Chieh Hung, and Marnie Collins (University of Melbourne). CSL Ltd. is the industry partner of the CRC for Vaccine Technology.
REFERENCES
Anonymous. 2002. The cost of antibiotic resistance: effect of resistance among Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa on length of hospital stay. Infect. Control Hosp. Epidemiol. 23:106-108.
Bell, J. M., J. D. Turnidge, A. C. Gales, M. A. Pfaller, and R. N. Jones. 2002. Prevalence of extended spectrum beta-lactamase (ESBL)-producing clinical isolates in the Asia-Pacific region and South Africa: regional results from SENTRY Antimicrobial Surveillance Program (1998-1999). Diagn. Microbiol. Infect. Dis. 42:193-198.
Blanchette, E. A., and S. J. Rubin. 1980. Seroepidemiology of clinical isolates of Klebsiella in Connecticut. J. Clin. Microbiol. 11:474-478.
Brooks, S. E., M. A. Walczak, S. Malcolm, and R. Hameed. 2004. Intrinsic Klebsiella pneumoniae contamination of liquid germicidal hand soap containing chlorhexidine. Infect. Control Hosp. Epidemiol. 25:883-885.
Butler, J. C., R. F. Breiman, H. B. Lipman, J. Hofmann, and R. R. Facklam. 1995. Serotype distribution of Streptococcus pneumoniae infections among preschool children in the United States, 1978-1994: implications for development of a conjugate vaccine. J. Infect. Dis. 171:885-889.
Butler, J. C., J. Hofmann, M. S. Cetron, J. A. Elliott, R. R. Facklam, and R. F. Breiman. 1996. The continued emergence of drug-resistant Streptococcus pneumoniae in the United States: an update from the Centers for Disease Control and Prevention's Pneumococcal Sentinel Surveillance System. J. Infect. Dis. 174:986-993.
Campbell, W. N., E. Hendrix, S. Cryz, Jr., and A. S. Cross. 1996. Immunogenicity of a 24-valent Klebsiella capsular polysaccharide vaccine and an eight-valent Pseudomonas O-polysaccharide conjugate vaccine administered to victims of acute trauma. Clin. Infect. Dis. 23:179-181.
Cryz, S. J., Jr., F. Furer, and R. Germanier. 1984. Experimental Klebsiella pneumoniae burn wound sepsis: role of capsular polysaccharide. Infect. Immun. 43:440-441.
Cryz, S. J., Jr., P. M. Mortimer, V. Mansfield, and R. Germanier. 1986. Seroepidemiology of Klebsiella bacteremic isolates and implications for vaccine development. J. Clin. Microbiol. 23:687-690.
Di Martino, P., Y. Bertin, J. P. Girardeau, V. Livrelli, B. Joly, and A. Darfeuille-Michaud. 1995. Molecular characterization and adhesive properties of CF29K, an adhesin of Klebsiella pneumoniae strains involved in nosocomial infections. Infect. Immun. 63:4336-4344.
Di Martino, P., N. Cafferini, B. Joly, and A. Darfeuille-Michaud. 2003. Klebsiella pneumoniae type 3 pili facilitate adherence and biofilm formation on abiotic surfaces. Res. Microbiol. 154:9-16.
Di Martino, P., V. Livrelli, D. Sirot, B. Joly, and A. Darfeuille-Michaud. 1996. A new fimbrial antigen harbored by CAZ-5/SHV-4-producing Klebsiella pneumoniae strains involved in nosocomial infections. Infect. Immun. 64:2266-2273.
Edelman, R., D. N. Taylor, S. S. Wasserman, J. B. McClain, A. S. Cross, J. C. Sadoff, J. U. Que, and S. J. Cryz. 1994. Phase 1 trial of a 24-valent Klebsiella capsular polysaccharide vaccine and an eight-valent Pseudomonas O-polysaccharide conjugate vaccine administered simultaneously. Vaccine 12:1288-1294.
Fang, C. T., Y. P. Chuang, C. T. Shun, S. C. Chang, and J. T. Wang. 2004. A novel virulence gene in Klebsiella pneumoniae strains causing primary liver abscess and septic metastatic complications. J. Exp. Med. 199:697-705.
Favre-Bonte, S., B. Joly, and C. Forestier. 1999. Consequences of reduction of Klebsiella pneumoniae capsule expression on interactions of this bacterium with epithelial cells. Infect. Immun. 67:554-561.
Fung, C. P., B. S. Hu, F. Y. Chang, S. C. Lee, B. I. Kuo, M. Ho, L. K. Siu, and C. Y. Liu. 2000. A 5-year study of the seroepidemiology of Klebsiella pneumoniae: high prevalence of capsular serotype K1 in Taiwan and implication for vaccine efficacy. J. Infect. Dis. 181:2075-2079.
Jagnow, J., and S. Clegg. 2003. Klebsiella pneumoniae MrkD-mediated biofilm formation on extracellular matrix- and collagen-coated surfaces. Microbiology 149:2397-2405.
Kaufmann, M. 1998. Pulsed-field gel electrophoresis, p. 33-50. In N. Woodford and A. Johnson (ed.), Methods in molecular medicine 15: molecular bacteriology: protocols and clinical applications. Humana Press, Totowa, N.J.
Legakis, N. J., L. S. Tzouvelekis, G. Hatzoudis, E. Tzelepi, A. Gourkou, T. L. Pitt, and A. C. Vatopoulos. 1995. Klebsiella pneumoniae infections in Greek hospitals. Dissemination of plasmids encoding an SHV-5 type beta-lactamase. J. Hosp. Infect. 31:177-187.
Livrelli, V., C. De Champs, P. Di Martino, A. Darfeuille-Michaud, C. Forestier, and B. Joly. 1996. Adhesive properties and antibiotic resistance of Klebsiella, Enterobacter, and Serratia clinical isolates involved in nosocomial infections. J. Clin. Microbiol. 34:1963-1969.
Maroncle, N., D. Balestrino, C. Rich, and C. Forestier. 2002. Identification of Klebsiella pneumoniae genes involved in intestinal colonization and adhesion using signature-tagged mutagenesis. Infect. Immun. 70:4729-4734.
Merchant, M., D. R. Karnad, and A. A. Kanbur. 1998. Incidence of nosocomial pneumonia in a medical intensive care unit and general medical ward patients in a public hospital in Bombay, India. J. Hosp. Infect. 39:143-148.
Nassif, X., J. M. Fournier, J. Arondel, and P. J. Sansonetti. 1989. Mucoid phenotype of Klebsiella pneumoniae is a plasmid-encoded virulence factor. Infect. Immun. 57:546-552.
Paganin, F., F. Lilienthal, A. Bourdin, N. Lugagne, F. Tixier, R. Genin, and J. L. Yvin. 2004. Severe community-acquired pneumonia: assessment of microbial aetiology as mortality factor. Eur. Respir. J. 24:779-785.
Palfreyman, J. M. 1978. Klebsiella serotyping by counter-current immunoelectrophoresis. J. Hyg. (London) 81:219-225.
Podschun, R., and U. Ullmann. 1998. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin. Microbiol. Rev. 11:589-603.
Schembri, M. A., D. Dalsgaard, and P. Klemm. 2004. Capsule shields the function of short bacterial adhesins. J. Bacteriol. 186:1249-1257.
Schurtz, T. A., D. B. Hornick, T. K. Korhonen, and S. Clegg. 1994. The type 3 fimbrial adhesin gene (mrkD) of Klebsiella species is not conserved among all fimbriate strains. Infect. Immun. 62:4186-4191.
Smith, S. M., J. T. Digori, and R. H. Eng. 1982. Epidemiology of Klebsiella antibiotic resistance and serotypes. J. Clin. Microbiol. 16:868-873.
Tambyah, P. A., V. Knasinski, and D. G. Maki. 2002. The direct costs of nosocomial catheter-associated urinary tract infection in the era of managed care. Infect. Control Hosp. Epidemiol. 23:27-31.
Tenover, F. C., R. D. Arbeit, R. V. Goering, P. A. Mickelsen, B. E. Murray, D. H. Persing, and B. Swaminathan. 1995. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J. Clin. Microbiol. 33:2233-2239.(Adam W. Jenney, Abigail C)