| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Clinical Microbiology, Aalborg Hospital, Aarhus University Hospital, Postbox 365, 9100 Aalborg, Denmark
2 Department of Clinical Epidemiology, Aalborg Hospital, Aarhus University Hospital, Aarhus, Denmark
3 The International Escherichia and Klebsiella Reference Center (WHO), Statens Serum Institut, Copenhagen, Denmark
4 Department of Clinical Microbiology, Hillerød Sygehus, Denmark
5 AdvanDx Inc., Woburn, MA, USA
Correspondence
Mette Søgaard
mette.soegaard{at}rn.dk
Received 12 July 2006
Accepted 5 March 2007
Abbreviations: BC, blood cultures; FISH, fluorescence in situ hybridization; PNA, peptide nucleic acid.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
After Escherichia coli, Klebsiella pneumoniae is the most common cause of Gram-negative bacteraemia and within its genus it accounts for three-quarters of all isolates, with some local variations (Feldman et al., 1990; Hansen et al., 1998). There is some controversy around the best way of classifying members of the genus Klebsiella, and as definitive identification by conventional methods may be cumbersome (Hansen et al., 2004), new methods for accelerated diagnosis would be advantageous.
Fluorescence in situ hybridization (FISH) with peptide nucleic acid (PNA) probes is a novel technique using fluorescein-labelled PNA probes targeting rRNA genes for rapid diagnosis of major BC pathogens (Oliveira et al., 2002; Rigby et al., 2002; Søgaard et al., 2005). The objectives of this study were to (i) evaluate a new PNA probe targeting K. pneumoniae using a panel of well-characterized strains representing Klebsiella spp. and the closely related Enterobacter aerogenes, and (ii) determine the utility of this assay for rapid identification of K. pneumoniae in routinely processed BCs.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Type strains and well-characterized clinical isolates. Type strains for the following species and subspecies were included in the evaluation: ATCC 13883 (K. pneumoniae), NCTC 5050 (Klebsiella ozaenae), ATCC 13182 (Klebsiella oxytoca), ATCC 33257 (Klebsiella terrigena), ATCC 33531 (Klebsiella planticola), JCM 6096 (Klebsiella ornithinolytica) and NCTC 10006 (Enterobacter cloacae). We also obtained the following panel of well-characterized clinical isolates from the collection at the International Escherichia and Klebsiella Reference Center (WHO) (Statens Serum Institut, Copenhagen, Denmark): K. pneumoniae (n=5), K. ozaenae (n=1), K. oxytoca (n=5), Klebsiella variicola (n=1), K. terrigena (n=5), K. planticola (n=5), K. ornithinolytica (n=5) and Enterobacter cloacae (n=5). Strains were biochemically characterized (Hansen et al., 2004) and additional sequencing of the 16S rRNA gene or the gyrA gene was performed (Boye & Hansen, 2003; Brisse & Verhoef, 2001).
All strains were grown by inoculating 0.1 ml of a 103–104 c.f.u. ml–1 bacterial suspension in routine aerobic FAN BC bottles with a negative growth index after 7 days' incubation. Bottles were loaded promptly into the BacT/Alert BC system and incubated until flagged positive and then unloaded and handled as described for the K. pneumoniae PNA FISH method. Reculture and retesting was performed for all positive reactions and six randomly chosen strains with negative reactions; otherwise strains were tested only once.
BC isolates. A BC set comprised one standard aerobic bottle, one FAN aerobic bottle and one anaerobic standard bottle for adult patients. An aerobic paediatric FAN bottle was used for children under the age of 6 years. Positive BCs were examined by direct microscopy and Gram staining, and cultures with Gram-negative rods were included as long as only one morphological type was observed. Only one bottle per BC set was included; if two or three bottles in a set were positive, the FAN bottle was selected (n=166); otherwise the anaerobic bottle (n=50) or the standard aerobic bottle (n=38) was selected, in that order. Ten negative BCs were included in the analysis as negative controls.
A total of 264 BCs with Gram-negative rods were included: K. pneumoniae (including one K. pneumoniae subsp. ozaenae), 80; K. oxytoca, 35; K. planticola, 1; Escherichia coli, 34; Citrobacter spp., 8; Enterobacter spp., 14; Morganella morganii, 5; Proteus spp., 11; Serratia spp., 6; Yersinia sp., 1; Salmonella spp., 9; Moraxella sp., 1; Capnocytophaga sp., 1; Haemophilus spp., 11; Acinetobacter spp., 7; Agrobacterium spp., 2; Burkholderia sp., 1; Pseudomonas spp., 16; Stenotrophomonas maltophilia, 2; unidentified aerobic Gram-negative rod, 1; Bacteroides spp., 15; Fusobacterium sp., 2; unidentified anaerobic Gram-negative rod, 1.
PNA FISH. The PNA probe was synthesized by Applied Biosystems. The probe sequence (5'-Flu-OO-CACCTACACACCAGC-3', where Flu=5,6-carboxyfluorescein and O=8-amino-3,6-dioxaoctanoic acid) is directed against a region of the 23S rRNA gene of K. pneumoniae starting at nt 1704 (Ludwig et al., 1995). The probe was designed to take advantage of a region of sequence heterogeneity between K. pneumoniae and K. oxytoca 23S rRNA genes, similar to a previously published DNA probe (Kempf et al., 2000).
The PNA FISH assay was performed according to the manufacturer's instructions (Oliveira et al., 2002). In brief, 10 µl BC broth was mixed on one-well Teflon-coated microscope slides (AdvanDx) with a drop of fixation solution, air-dried and fixed by flame fixation. Slides not examined immediately were stored at room temperature in the dark. One drop of hybridization solution containing the fluorescein-labelled PNA probe was added and a coverslip was applied. For hybridization, slides were incubated on a slide warmer (SM30; Grant Boekel) at 55 °C for 90 min. Subsequently, coverslips were removed and the slides were submerged for 30 min in a pre-heated wash solution (55 °C) in a water bath. Slides were mounted with mounting fluid (KT001; AdvanDx) and coverslips were applied before examination under a fluorescence microscope (x100 objective, Olympus BX40; Osram HBO 100 W/2 Hg lamp) equipped with a fluorescein isothiocyanate-–Texas red dual band-pass filter (filter no. AC003; AdvanDx).
Two observers who were blind to the results by conventional identification examined the slides. Distinct green fluorescence of multiple bacteria was scored as a positive result and graded as bright or faint. Negative slides had a faintly reddish background; rarely, a few fluorescent bacteria or bacterial clusters were observed, but lower fluorescence intensity and cell numbers clearly distinguished these from the positive results.
Conventional identification. We performed subculture on 5 % horse blood agar, chocolate agar, CPS2 agar (bioMérieux) and K vitamin-enriched chocolate agar for anaerobic culture, as deemed appropriate. Blood culture isolates were identified tentatively on the basis of colony morphology and Gram staining. The CPS2 agar facilitated diagnostic shortcuts for typical Escherichia coli (ß-glucuronidase-positive, indicated by burgundy-coloured colonies, and spot indole-positive) and swarming Proteus spp. (tryptophan deaminase-positive and spot indole-positive). All other enterobacteria were identified with the ID-GNB card in the Vitek2 system (bioMérieux); a negative spot indole test was required for the identification of K. pneumoniae.
The results obtained by conventional identification were used as reference. If a discrepancy between conventional identification and PNA FISH was observed, the isolate was retested by use of simulated culture and referred to the International Escherichia and Klebsiella Reference Center (WHO) (Statens Serum Institut) for extensive biochemical testing (Hansen et al., 2004).
The performance characteristics of the K. pneumoniae probe were evaluated by calculation of the sensitivity, specificity and positive and negative predictive values.
|
RESULTS AND DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
We included Enterobacter aerogenes (a valid synonym of K. mobilis), as this species, based on phenotypic traits and DNA relatedness, is more closely related to Klebsiella spp. than to Enterobacter cloacae (the type species of the genus Enterobacter) (Bascomb et al., 1971; see also discussion by Boye & Hansen, 2003). K. variicola (‘adonitol non-fermenting K. pneumoniae’) appears to correspond to the KpIII cluster of K. pneumoniae (Brisse et al., 2004) and we therefore regard both K. ozaenae and K. variicola as members of the K. pneumoniae complex.
Among the species included, only K. pneumoniae, K. oxytoca and Enterobacter aerogenes are frequently reported BC isolates. The lack of easily determinable and distinct phenotypic characteristics within the genus Klebsiella is a limitation of commercial identification systems and, as a consequence, the clinical importance of other Klebsiella spp. is known only sporadically (Liu et al., 1997; Podschun et al., 2001; Rosenblueth et al., 2004; Westbrook et al., 2000). Seen in this perspective, easy identification of K. pneumoniae, K. ozaenae and K. variicola by PNA FISH could be helpful.
Diagnostic performance
Seventy-nine out of 82 Gram-negative rods conventionally identified as K. pneumoniae were positive by PNA FISH and three were negative. When re-examined by extensive biochemical tests, two of the three isolates were found to be K. oxytoca and K. planticola. The last false negative yielded a bright positive result upon retesting with PNA FISH.
Among 184 Gram-negative rods not identified as K. pneumoniae, 182 were negative by PNA FISH and two were positive. One discrepant culture, which stained only faintly and was negative by PNA FISH when retested, was identified as Pseudomonas aeruginosa. The other discrepant culture was initially identified as Klebsiella sp., but further biochemical testing showed it to be K. pneumoniae. Thus we obtained only one false-positive result.
Based on these data, the performance specifications for the K. pneumoniae PNA FISH assay were: diagnostic sensitivity 98.8 % (79/80 specimens), diagnostic specificity 99.5 % (183/184 specimens), positive predictive value 98.8 % (79/80 specimens) and negative predictive value 99.5 % (183/184 specimens). These estimates are likely to be conservative because of the low proportion of Escherichia coli isolates included in our study and because all Escherichia coli isolates were uniformly negative.
Except for this under-representation of Escherichia coli, the study was conducted in a routine setting and the few spurious results that occurred might have been due to handling problems in the laboratory. The preparatory steps prior to hybridization could be critical, as a false-negative K. pneumoniae isolate was positive upon retest and the false-positive P. aeruginosa isolate was negative when retested.
Our results corroborate previous studies with PNA FISH assays targeting other microbial pathogens (Oliveira et al., 2002; Rigby et al., 2002; Søgaard et al., 2005). PNA probes are known to have an overall superior performance to DNA probes when used in the FISH format, including greater specificity (Wilks & Keevil, 2006). The absence of complete sequence information for the rarer Klebsiella spp. argued for selection of the more-specific (PNA) probes in this study.
K. pneumoniae, Escherichia coli and P. aeruginosa are the major Gram-negative BC pathogens and the combined use of probes targeting these would allow rapid and specific identification of nearly two-thirds of Gram-negative bacteraemias. In countries like Denmark, where most bacterial groups still have a predictable pattern of resistance, such a diagnostic strategy could support early directed antibiotic therapy.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
