| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratory Medicine Services, Changi General Hospital, Singapore
Correspondence
Thean Yen Tan
thean_yen_tan{at}cgh.com.sg
Received 5 May 2006
Accepted 20 November 2006
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Identification of Gram-negative bacilli at least to the genus level is required for the proper interpretation of susceptibility test results. Identification of the infecting bacterium may assist clinicians in determining the source of sepsis and initiating further investigations.
Multiple techniques have been described for the identification of Gram-negative bacteria from positive blood cultures. These include the direct inoculation of blood-broth aspirates onto Vitek test cards (de Cueto et al., 2004), chromogenic or fluorogenic substrates (Sepulveda et al., 1990), enzymic test strips (Edberg et al., 1975) and commercial biochemical identification kits (Malloy et al., 1983). Molecular testing methods have utilized fluorescent hybridization probes (Kempf et al., 2000) and amplification of 16S rRNA genes from positive blood culture vials (Qian et al., 2001).
Clinical microbiology laboratories face increasing manpower pressure and cost constraints (Peterson et al., 2001). Most published methods either involve the use of expensive consumables or are too labour-intensive for routine implementation. In this study, we tested the accuracy of an identification system for members of the Enterobacteriaceae and Acinetobacter spp. based on the direct inoculation of conventional biochemical tests directly from positive blood culture vials.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Identification of Gram-negative bacilli. Broth from the positive vial was subcultured directly onto trypticase soy agar with 5 % sheep blood and onto MacConkey agar. A 0.5 ml sample of the broth was diluted in 2.5 ml sterile 0.85 % saline. This blood/saline dilution was used to inoculate a set of commercial biochemical tubes. The tests used were a Kligler iron agar slant, a Simmons' citrate agar slant, a urea hydrolysis slant, a phenylalanine deaminase agar slant, tryptophan broth, malonate broth and an open tube of oxidative–fermentative medium. All media were incubated overnight at 35 °C under ambient atmospheric conditions.
Following incubation, specimens with more than one colonial morphotype of Gram-negative bacilli were excluded from the study. Isolates were tested for oxidase production using a commercial reagent (Becton Dickinson) and oxidase-positive Gram-negative bacilli were excluded from further analysis.
Biochemical tube reactions were read and interpreted according to recommended guidelines (MacFaddin, 2000). The genus of the isolate was determined according to the criteria in Table 1
(derived from data in Farmer, 2003).
|
The genus identifications obtained by the study method and commercial identification systems were compared. Confirmatory biochemical identification by Vitek II using ID-GN panels (bioMérieux) was used to resolve study isolates with discrepant genus identification. The final genus identification was accepted if the genus identification obtained by the study method and by the initial commercial kits was concordant or if two out of the three identification systems provided similar genus identification.
Labour, times and costing. The time used to inoculate biochemical tests for the study method by different technologists was recorded on multiple occasions using a stopwatch. Test timings included the labelling and inoculation of media, interpretation of biochemical test reactions and the recording of results onto worksheets. Timing results were analysed using Microsoft EXCEL. Statistical analysis of means and standard deviation was performed using EXCEL.
The time taken to generate genus identification was calculated from the initial processing of a blood culture vial with a positive growth index until the provision of genus identification. The duration of overnight incubation was therefore included in the timing calculations.
The cost of consumables was based on prices in Singapore dollars (SGD) and converted to the British pound at the prevailing exchange rate of SGD1=£0.34.
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Genus identification was concordant with the initial testing by commercial identification systems in 131 out of the 136 isolates (96 %). The five isolates with discrepant results underwent confirmatory testing by Vitek II, following which a final accuracy of 98 % for the study method (133/136 isolates) was achieved. The accuracy of genus identification for the direct testing method when compared with the reference methods is detailed in Table 2.
|
Labour, timings and costing
The mean time required to inoculate a single biochemical test from a positive blood culture vial was 22 s (standard deviation 5, range 18–36), whilst interpretation and recording of results took 7 s per test (standard deviation 1, range 6–8) (Fig. 1). The total mean time required for inoculation of the full test panel, interpretation and recording of results was 3 min per positive blood culture vial.
|
The comparative costs of consumables including the use of test reagents were: biochemical tube tests, £2.40; API 20E, £5.14; and Vitek ID-GN panel, £9.11.
|
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study did not measure the clinical impact of providing faster identification of Gram-negative bacilli from blood cultures. A study by Trenholme et al. (1989) demonstrated that provision of direct testing results from positive blood cultures was more likely to result in changes in antibiotic therapy. Antibiotic susceptibilities of Gram-negative bacteria should be interpreted concomitant with the genus identification, as this allows improved interpretation of antibiotic susceptibilities (Livermore et al., 2001). The use of a direct inoculation method to provide genus identification may be combined with direct antibiotic susceptibility testing for earlier and more accurate reporting of positive blood cultures. Direct antibiotic susceptibility testing from positive blood culture vials has been advocated as a means of providing rapid results (Baron et al., 2005), which has been shown to be associated with clinical and financial benefits (Barenfanger et al., 1999).
The study method is less labour-intensive than other previously published methods, which usually require an additional centrifugation step (de Cueto et al., 2004; Waites et al., 1998), and may fit more easily into the workflow of a clinical microbiology laboratory.
We were unable to compare directly the full costs of the study method with those of previously published methods of direct bacterial identification from positive blood cultures, as data were lacking in the earlier studies on technologist time required to perform the reported tests. However, the costs of consumables for this method are lower than alternative methods that require the use of commercial identification systems. Financial considerations for direct testing methods are important because blood cultures with more than one bacterial species comprise 7–17 % of all positive blood cultures (Ispahani et al., 1987; Pittet et al., 1997). Other studies have reported that 7–16 % of positive blood cultures with a single bacterial morphology on Gram staining are polymicrobial (Moore et al., 1981; Waites et al., 1998). In our study, 25 % of blood cultures that were tested initially by the direct method had to be excluded subsequently because of mixed bacterial morphotypes following overnight culture.
The study method is only suitable for the identification of common clinical isolates of members of the Enterobacteriaceae and Acinetobacter species. Non-glucose-fermenting Gram-negative bacilli, fastidious organisms such as Haemophilus influenzae and oxidase-positive glucose fermenters such as Aeromonas spp. are not identifiable by this method. E. coli and Klebsiella spp. formed the predominant isolates recovered in this study, with smaller numbers of other genera of the Enterobacteriaceae tested. Further validation of the study method may be required for bacteria other than E. coli and Klebsiella spp.
There are several limitations to the study method. Differentiation between Klebsiella spp. and Enterobacter spp. was based solely on the motility of the test isolates. Weakly motile organisms may not be visualized clearly in oxidative–fermentative tubes. Furthermore, not all Enterobacter spp. are motile (Abbott, 2003) and as such may be misidentified as Klebsiella spp. The addition of an ornithine decarboxylation test to the biochemical test series, although not performed in this study, would enable better differentiation between the two genera. Determination of a bacterium's ability to utilize citrate as a sole source of carbon energy requires >24 h incubation. This may inherently result in the misinterpretation of Providencia rettgeri as M. morganii when the test is interpreted as negative following overnight incubation, as suggested in this study. Hydrogen sulfide-producing, non-lactose-fermenting C. freundii may only be differentiated from Salmonella spp. by the absence of Salmonella antigens. Although some isolates in the study were identified to species level, the study method primarily identifies Gram-negative bacilli to genus level. The accuracy of the study method is not equivalent to that obtained by commercial identification systems.
This study provides data for a method of performing direct genus identification of Gram-negative bacilli from positive blood culture vials, with acceptable levels of accuracy. Implementation of this method with concurrent direct antibiotic susceptibility testing allows for accurate interpretation of antibiotic susceptibility results, allowing provisional diagnosis and treatment to be made whilst allowing time for confirmatory test results by conventional methods. This method is suitable for routine clinical use, as it is comparatively less costly than other proposed methods and is not labour-intensive.
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
