Evaluation of the BD Phoenix Automated Microbiology System for Identification and Antimicrobial Susceptibility Testing of Staphylococci and
http://www.100md.com
微生物临床杂志 2006年第6期
Division of Microbiology, Department of Pathology, the Johns Hopkins University School of Medicine
Microbiology Laboratory, the Johns Hopkins Hospital, Baltimore, Maryland 21287
ABSTRACT
We evaluated the Phoenix automated microbiology system (BD Diagnostic Systems, Sparks, MD) for the identification (ID) and antimicrobial susceptibility testing (AST) of challenge and clinical staphylococci and enterococci recovered from patients in a tertiary-care medical center. In total, 424 isolates were tested: 90 enterococci; 232 Staphylococcus aureus isolates, including 14 vancomycin-intermediate S. aureus isolates; and 102 staphylococci other than S. aureus (non-S. aureus). The Phoenix panels were inoculated according to the manufacturer's instructions. The reference methods for ID comparisons were conventional biochemicals and cell wall fatty acid analysis with the Sherlock microbial identification system (v 3.1; MIDI, Inc. Newark, DE). Agar dilution was the reference AST method. The overall rates of agreement for identification to the genus and the species levels were 99.7% and 99.3%, respectively. All S. aureus isolates and enterococci were correctly identified by the Phoenix panels. For the non-S. aureus staphylococci, there was 98.0% agreement for the ID of 16 different species. The AST results were stratified by organism group. For S. aureus, the categorical agreement (CA) and essential agreement (EA) were 98.2% and 98.8%, respectively. Three of three very major errors (VMEs; 1.7%) were with oxacillin. For non-S. aureus staphylococci, the CA, EA, VME, major errors, and minor error rates were 95.7%, 96.8%, 0.7%, 1.7%, and 2.9%, respectively. The two VMEs were with oxacillin. For the enterococci, there was 100% CA and 99.3% EA. All 36 vancomycin-resistant enterococci were detected by the Phoenix system. The Phoenix system compares favorably to traditional methods for the ID and AST of staphylococci and enterococci.
INTRODUCTION
The burden of the rapid and accurate detection of antimicrobial resistance among gram-positive bacteria remains a continuous challenge for clinical microbiology laboratories. The last decade has seen increases in the numbers of vancomycin-resistant enterococcal (VRE) infections (11, 18), the appearance of glycopeptide resistance among staphylococci (15, 18), a continued rise in nosocomial methicillin-resistant Staphylococcus aureus (MRSA) isolates (9, 11), and a surge in community-associated MRSA isolates (5, 12, 19). In addition, other staphylococci continue to cause serious nosocomial infections, such as endocarditis caused by S. lugdunensis and infections of intravascular and prosthetic devices caused primarily by S. epidermidis (1, 7, 13). This fact, combined with the change in breakpoints for oxacillin/methicillin testing of coagulase-negative staphylococci (10), has presented additional challenges for laboratories, which must adopt systems with improved abilities for the identification (ID) and antimicrobial susceptibility testing (AST) of non-S. aureus staphylococci.
The newer automated instruments have more extensive databases, data management tools (including expert systems), and other features unique to the instrument of each manufacturer. In addition, the accuracy of susceptibility testing has improved with these systems, although problems with the detection of some resistance phenotypes remain (e.g., glycopeptide resistance among S. aureus), leading to recommendations for the continued use of manual screening methods (8, 15, 18).
The newest system to obtain FDA clearance is the Phoenix automated microbiology system (BD Diagnostic Systems, Sparks, MD). This automated system identifies a broad range of gram-positive and gram-negative bacteria. The system includes an instrument, software, disposable panels, broths for ID and AST, and a susceptibility testing indicator. The ID method uses modified conventional, fluorogenic, and chromogenic substrates. The instrument can analyze up to 100 ID and AST combination panels simultaneously. The disposable test panels contain 136 microdilution wells and are available in ID-only, ID/AST, and AST-only formats. The panels are read every 20 min by the instrument.
This study evaluated the performance of the Phoenix instrument for the identification and susceptibility testing of challenge and clinical staphylococci and enterococci isolated from a variety of specimen sources.
MATERIALS AND METHODS
Bacterial strains. One hundred sixty-two previously characterized "challenge" gram-positive cocci and 262 clinical isolates prospectively recovered from routine cultures in the Clinical Microbiology Laboratory from the Johns Hopkins Hospital (JHH) were included in this comparison (the total number of isolates tested was 424). The challenge isolates included 107 S. aureus isolates, including 14 vancomycin-intermediate S. aureus strains (VISA) and 79 clinical S. aureus isolates that were characterized by staphylococcal cassette chromosome mec typing (SCCmec), Panton-Valentine leukocidin (PVL) gene testing, and pulsed-field gel electrophoresis (PFGE) by previously published methods (4). The 79 S. aureus clinical isolates were obtained during an epidemiological study of community-associated MRSA. Also among the challenge isolates were 6 enterococci and 49 coagulase-negative staphylococci that were previously either difficult to identify or that previously had susceptibility patterns that had required supplemental testing. The non-S. aureus challenge isolates included the following: S. epidermidis ATCC 14990, S. saprophyticus ATCC 15305, S. cohnii ATCC 49330, S. warneri ATCC 27836, S. capitis subsp. capitis ATCC 49324, S. chromogenes ATCC 43764, S. intermedius ATCC 29663, S. equorum ATCC 43958, S. gallinarum ATCC 35539, S. kloosii, ATCC 43959, and Macrococcus caseolyticus ATCC 13548.
The prospective clinical isolates included 125 S. aureus isolates, 53 coagulase-negative staphylococci, and 84 enterococcal species. All organisms except for the 14 VISA isolates were from unique patients. Eleven VISA isolates (MICs, 4 to 16 μg/ml) were recovered over time from the same patient who had been treated with prolonged antibiotics for osteomyelitis. In addition, three strains of VISA (MICs, 4 to 8 μg/ml by broth dilution) from wounds or tissues from two sites within the United States and one from Brazil (designated strains NRS 1, NRS 56, and NRS 73, respectively) were obtained through the Network on Antimicrobial Resistance in Staphylococcus aureus (supported under NIAID, NIH, contract no. N01-AI-95359). These VISA isolates are not included in the tables of cumulative data but are addressed separately in the Results section.
Reference identification. The laboratory's routine method for the identification of gram-positive organisms includes testing by the following conventional rapid tests and by assays with biochemicals incorporated into agar media. For S. aureus, slide coagulase and exogenous nuclease tests were routinely performed. If needed, the tube coagulase test, polymyxin B susceptibility testing, a test for ornithine decarboxylation, and a test for fermentation of mannitol were additionally performed. All reactions with the exception of the slide coagulase reaction were read at 24 h. For species-level identification of coagulase-negative staphylococci, a combination of conventional biochemical tests and cell wall fatty acid analysis was performed. The biochemical assays performed included tests for fermentation of sucrose, lactose, mannitol, arabinose, turanose, trehalose, and mannose; a urease detection test; and novobiocin and polymyxin B susceptibility testing. Cellular fatty acid analysis was performed by using the Sherlock microbial identification system (MIDI, Newark, DE). Organism identification was based on computer comparison of the unknown organism's fatty acid methyl ester profile with predetermined fatty acid methyl ester profiles in the library of the microbial identification system with software version 3.1. Sugar fermentation was determined by using a peptone agar base with phenol red indicator. All reactions were interpreted according to published guidelines (1). The enterococci were tested for the following: bile esculin; 6.5% sodium chloride; motility; colony pigmentation; and fermentation of sucrose, lactose, mannitol, sorbitol, arabinose, and sorbose.
Phoenix system identification method. The Phoenix identification method uses modified conventional, fluorogenic, and chromogenic substrates. Combination panels for investigational use only (PMIC/ID-33, catalog no. 448587) for both identification and susceptibility testing were used for this comparison. Software V3.34A/V3.54A was used for this study. The ID side contains 45 wells with dried biochemical substrates and 2 fluorescent control wells. The ID broth was inoculated with bacterial colonies adjusted to a 0.5 McFarland standard by using a CrystalSpec nephelometer (BD Diagnostics), according to the manufacturer's recommendations. The suspension was then poured into the ID side of the Phoenix panel after an aliquot (25 μl) was removed for AST. The specimen was logged and loaded into the instrument within the specified timeline of 30 min. Quality control and maintenance were performed according to the manufacturer's recommendations.
Reference AST. Agar dilution was performed with plates prepared in-house according to CLSI (formerly NCCLS) guidelines (10). The following antibiotics were tested at the indicated concentrations: penicillin, 0.1 to 8 μg/ml; oxacillin, 0.125 to 2 μg/ml; erythromycin, 0.5 to 4 μg/ml; clindamycin, 0.5 to 2 μg/ml; ampicillin, 2 to 16 μg/ml; trimethoprim-sulfamethoxazole, 0.5 and 2 μg/ml; gatifloxacin, 0.5 to 4 μg/ml; vancomycin, 1 to 16 μg/ml; and nitrofurantoin, 32 and 64 μg/ml (urine samples only). For enterococci, a screen for high-level aminoglycoside resistance was also included. MecA gene testing with the LightCycler instrument (Roche Molecular Systems, Pleasanton, CA) was used to confirm any discrepant results for oxacillin obtained by the agar method and with the Phoenix instrument. To determine whether the discrepancies for gatifloxacin were related to methodological differences between an agar-based susceptibility testing method and a broth-based system (as has been reported for other quinolones and members of the family Enterobacteriaceae) (17), broth macrodilution was performed according to CLSI guidelines.
Phoenix system AST. The AST side of the combination panel contains 84 wells with dried antimicrobial panels and one growth control well. The assay is a broth-based microdilution test. The system uses a redox (AST) indicator for the detection of organism growth in the presence of an antimicrobial agent. One free-falling drop of the AST indicator was added to the AST broth tube. Twenty-five microliters of the standardized ID broth suspension was transferred to the AST broth, yielding a final concentration of approximately 5 x 105 CFU/ml. Quality control was performed according to the manufacturer's recommendations. The following antibiotics were tested at the indicated concentrations, and the results were compared to those obtained by agar dilution: penicillin, 0.125 to 16 μg/ml; oxacillin, 0.125 to 4 μg/ml; erythromycin, 0.25 to 4 μg/ml; clindamycin, 0.25 to 4 μg/ml; ampicillin, 0.25 to 32 μg/ml; trimethoprim-sulfamethoxazole, 0.5/9.5 to 2/38 μg/ml; gatifloxacin, 1 to 4 μg/ml; vancomycin, 1 to 16 μg/ml; and nitrofurantoin, 16 to 64 μg/ml (urine samples only). During the study, the version of software that was used had a limitation with the results for coagulase-negative staphylococci (except S. epidermidis) and penicillin. Therefore, fewer comparative results are available for penicillin and coagulase-negative staphylococci. For enterococci, a screen for high-level aminoglycoside resistance for gentamicin and streptomycin was also included. For enterococci, only the results for ampicillin at 0.125 to 32 μg/ml, vancomycin at 1 to 16 μg/ml, tetracycline at 0.5 to 8 μg/ml, and nitrofurantoin at 16 to 64 μg/ml were considered in the data analysis.
Data analysis. All data from both the reference methods and the Phoenix instrument were entered into and analyzed by using Excel spreadsheets. For the ID portion of the test, accuracies to the genus and the species levels were determined. For AST essential agreement (EA) and categorical agreement (CA) were determined. Essential agreement was defined as MICs between the two systems that were within plus or minus 1 doubling dilution. Categorical agreement was defined as susceptible, intermediate, and resistant results that matched between the two systems. The rates of very major errors (VMEs; false-susceptible result with the Phoenix system), major errors (MEs; false-resistant result with the Phoenix system), and minor errors (mEs; one system reported an intermediate result while the other method reported a resistant or susceptible result) were calculated. The number of resistant strains was used as the denominator for the calculation of the VME rate. For the calculation of the ME rates, the number of susceptible strains was used as the denominator.
Discrepant resolution. For identification of discrepancies between the two systems, testing was repeated by both methods. Biochemical testing plus cellular fatty acid analysis by gas-liquid chromatography (GLC) was accepted as the "gold standard." For susceptibility testing, for any organism-drug combination with discrepant results between the two systems, the combination was retested by both methods. Agar dilution testing was accepted as the gold standard for all antimicrobials except oxacillin/methicillin. MecA gene testing was used to resolve the discrepancies obtained by oxacillin/methicillin testing. As stated above, broth macrodilution was used to resolve discrepancies related to gatifloxacin testing.
RESULTS
Table 1 presents the overall results for the identification of the staphylococcal and enterococcal species tested. Initially, 10 isolates (9 non-S. aureus staphylococci and a Macrococcus sp.) had discrepant results between the Phoenix system and conventional methods. Seven of these were resolved by repeat testing with both systems. The overall rates of agreement to the genus and the species levels were 99.7% and 99.3%, respectively, after repeat testing.
As stated above, there were three discrepant results. A Staphylococcus hominis isolate was misidentified by the Phoenix system as a "No ID" initially and then "S. capitis subsp. capitis" on repeat testing, and a Staphylococcus capitis subsp. capitis isolate was misidentified as Staphylococcus hemolyticus. A third discrepant result was obtained for Macrococcus caseolyticus, which was misidentified as S. kloosii. No strains of S. aureus were misidentified as coagulase-negative staphylococci, nor was the opposite true. All Staphylococcus epidermidis isolates and the three isolates of Staphylococcus lugdunensis were correctly identified.
For enterococcal identification, there was 100% agreement to the species level for the 90 strains of E. faecium and E. faecalis. The Phoenix system does not distinguish between E. casseliflavus and E. gallinarum, but it correctly assigned the two organisms in this study to the overlap category. Two E. raffinosus isolates were also correctly identified by the Phoenix instrument.
On initial susceptibility testing, 52 of the 410 clinical and challenge isolates had an organism-drug discrepancy by one of the two systems. After repeat testing, 30 of the isolates gave concordant results by both systems. After repeat testing, the overall EAs for staphylococci and enterococci were 98.2% and 99.3%, respectively. For CA the values were 97.4% and 100% for staphylococci and enterococci, respectively. The overall VME, ME, and mE rates for all organisms were 0.4%, 0.8%, and 1.7%, respectively.
Tables 2 to 4 stratify the susceptibility results by organism group. Table 2 summarizes the results for the 218 S. aureus strains. The overall EA and CA for S. aureus were 98.8% and 98.2%, respectively. Seventy-eight clinical MRSA isolates were recovered during the prospective portion of the study. Within this group, four isolates that tested resistant by the agar dilution method were called susceptible by the Phoenix system. mecA gene PCR confirmed that three of the four isolates were true MRSA isolates. Of these three isolates, the Phoenix system flagged one for additional testing because of its resistance phenotype. However, for two strains that lacked a resistance profile, initial results were susceptible for oxacillin, and the expert system failed to flag these isolates as requiring supplemental testing.
To further challenge the system, a group of 79 diverse MRSA isolates characterized by SCCmec typing, PCR for the Panton-Valentine leukocidin genes, and PFGE were tested. Forty-nine of the isolates were SCCmec type IV, of which 25 were PVL positive; and 23 of the isolates were SCCmec type II, of which none were PVL positive. For seven of the isolates, the SCCmec type could not be determined. The majority of the PVL-positive, SCCmec type IV strains belonged to the USA 300 clone. The Phoenix system called one strain from this group that was a resistant SCCmec type IV strain oxacillin susceptible. When this group was analyzed separately, the EA and CA were 99.6% and 97.8%, respectively (data not shown). When the isolates were stratified by SCCmec type and PFGE patterns, there were no significant differences in the ability of the Phoenix system to accurately identify or perform AST between community-associated MRSA isolates and health care-associated MRSA isolates (data not shown).
Overall, the VME, ME, and mE rates for the total number of S. aureus isolates were 0.4%, 0.3%, and 1.5%, respectively. The three MEs were seen with trimethoprim-sulfamethoxazole. All but four of the mEs were seen with gatifloxacin. Twenty-two strains showed discrepant results for gatifloxacin between the Phoenix system and the agar reference method. Of these 22 strains, 21 of them were intermediate by agar dilution, with an MIC of 4 μg/ml, whereas the Phoenix system called 19 of them susceptible (MIC, 2 μg/ml) and the remaining 2 resistant (MIC, >4 μg/ml). Twenty-one of the strains were retested by broth macrodilution to assess whether the discrepancies were related to methodological differences, as has been reported for other quinolones when other species are tested (17). Fourteen of the 18 (78%) available strains with discrepant intermediate-susceptible results had identical MICs of 2 μg/ml, like the Phoenix results, suggesting that methodological differences exist between agar dilution and broth dilution methods.
Fourteen isolates of MRSA with intermediate resistance to vancomycin by agar dilution (MICs, 4 to 16 μg/ml) were tested. Eleven isolates were recovered over a period of several weeks from cultures of bone and wound specimens from the same JHH patient with MRSA osteomyelitis who had received prolonged vancomycin treatment. The other three isolates were obtained from the Network on Antimicrobial Resistance in Staphylococcus aureus. The Phoenix system MICs for 10 of the JHH strains were 8 μg/ml, and for 1 of the strains the MIC was 16 μg/ml. The MICs of vancomycin for strains NRS 1 and NRS 56 were 8 μg/ml, and for NRS 73 the vancomycin MIC was 4 μg/ml. Based upon these MICs, the Phoenix expert system flagged these isolates as being unusual, suggesting the need for additional tests and, if the finding was verified, notification of infection control personnel. Data for these isolates have been excluded from the cumulative data in Tables 1 and 2.
Table 3 summarizes the findings for the non-S. aureus staphylococci. The overall EA and CA for this group were 96.8% and 95.7%, respectively. The VME, ME, and mE rates were 0.7%, 1.7%, and 2.9%. The two VMEs were seen with oxacillin, whereas trimethoprim-sulfamethoxazole accounted for the majority of the major errors. As was seen with S. aureus, eight of the minor errors were with gatifloxacin.
Table 4 lists the AST results for the 90 enterococcal strains tested. The overall EA and CA for enterococci were 99.3% and 100%, respectively. There were no errors for the enterococci. There were 36 VREs detected by the reference method and with the Phoenix instrument. Fourteen of these isolates were tested for high-level resistance to gentamicin by the reference method, and the result was compared to the Phoenix system's gentamicin high-level resistance result. Of the four isolates with high-level resistance to gentamicin as detected by the reference test, all were also detected by the Phoenix system.
DISCUSSION
The BD Phoenix-100 system is the newest automated instrument to be cleared and marketed by the FDA for use in the United States. This instrument has been available for some years in Europe, and literature on its performance for the identification and susceptibility testing of a broad range of microorganisms is beginning to accumulate. Perhaps of most concern to users is the ability of the system to correctly detect the broad range of resistance phenotypes that have become a universal concern to microbiologists. This study evaluated the reliability of the Phoenix system for the identification and detection of a large number of staphylococci and enterococci representing 23 species and the major resistance phenotypes, including nosocomial and community-associated MRSA clones, VISA, and VRE.
The Phoenix system compared favorably to classical biochemical tests and GLC-based methods for the identification of all of the staphylococci and enterococci tested. The isolates tested represent the strains of coagulase-negative staphylococci and enterococci most commonly recovered in our busy tertiary-care medical center (JHH) but are not inclusive of all possible clinically relevant species. In addition, a limitation of the identification portion of this study is the small numbers of some species available for testing during the evaluation period. The overall concordance rates for the isolates tested were 99.7% to the genus level and 99.3% to the species level. When the results are separated by genus, the Phoenix system had 99.7% and 100% accuracies for the identification of staphylococci and enterococci, respectively. These results are somewhat better than those observed by Donay et al. (2). In that study, there was a 91.8% concordance rate for Staphylococcus sp. identification and an 86.2% concordance rate for the identification of 29 Enterococcus spp. Discrepant results between the Phoenix system and the API 32 Staph identification system (bioMerieux, Inc., Durham, NC) were resolved by 16S rRNA gene sequencing. As in our study, 100% concordance was observed for S. aureus identification. In the two-center study by Fahr et al. (3), 275 staphylococci and 177 enterococci were tested. The rates of concordance were 97.1% and 98.9% for identification of these two groups, respectively, compared to the results obtained with the Vitek 2 and API systems (bioMerieux). None of the discordant results occurred with S. aureus. For the enterococci, two E. faecalis spp. were misidentified as E. casseliflavus-E. gallinarum (3). One hundred percent concordance for S. aureus identification was also seen in the study by Spanu et al. (14), where the overall concordance for the identification of staphylococci to the species level was 98.6% compared to the results obtained with the ID 32 Staph system.
In terms of susceptibility testing, we attempted to challenge the Phoenix with a variety of resistance phenotypes commonly seen in our laboratory environment. Unique to our study, compared to other studies reported in the literature, was the inclusion of a significant number of not only hospital-acquired isolates but also community-associated MRSA strains, as characterized by the presence of PVL genes, SCCmec typing, and PFGE.
The Phoenix system performed well for the detection of MRSA, both community-associated MRSA isolates and health care-associated MRSA isolates. The CA for oxacillin was 98.6%, and the EA for oxacillin was 95.4%. The rates of agreement between the agar dilution method and the Phoenix system were excellent for the other antibiotics evaluated. Fourteen VISA strains were correctly identified by the Phoenix expert system as requiring supplemental testing.
Minor errors were observed primarily for gatifloxacin. No publications have addressed the methodological differences between agar-based methods and broth-based methods for S. aureus when gatifloxacin is tested. Donay et al. (2) tested ofloxacin by agar diffusion and noted a 3.2% mE rate among non-S. aureus staphylococci with the Phoenix system, but this rate was not higher than the rates observed with several other antimicrobial agents. Methodological variations and the high mE rates obtained when members of the family Enterobacteriaceae were tested against two quinolones were reported by Steward et al. (17), who compared a reference broth microdilution method with an agar dilution method. Agar dilution had mE rates of 8.2% for ofloxacin and 12.3% for ciprofloxacin. Gatifloxacin was not tested in that evaluation (17). The limited data obtained by a broth macrodilution method in our study suggest that real methodological differences exist between the agar dilution and the broth dilution methods when staphylococci and gatifloxacin are tested. More studies are needed to confirm our results.
The Phoenix system was able to accurately detect oxacillin resistance among coagulase-negative staphylococci when agar dilution was used as the reference method. There was one VME. In the study by Horstkotte et al. (6), the Phoenix system had a 99.2% sensitivity for the detection of oxacillin/methicillin resistance at the current CLSI (formerly NCCLS) MIC breakpoint of 0.5 μg/ml compared to the results of mecA PCR, which was used as the reference method. By use of this breakpoint, the Phoenix system called 26 mecA-negative strains resistant (specificity, 64.9%). Similar results were obtained by a reference broth microdilution method. The authors concluded that the Phoenix system results were equivalent to those of other phenotypic methods but that confirmation of resistance by mecA PCR should be considered for isolates with oxacillin MICs between 0.5 and 2.0 μg/ml (6).
For enterococcal susceptibility testing, the Phoenix system performed well compared to the performance of the agar dilution method. Vancomycin resistance and -lactam resistance were accurately detected by the instrument. There were no VME or ME for any of the resistance profiles. In addition, high-level resistance to gentamicin was also accurately detected. These results compare favorably to those reported previously by others (2, 3, 16). These data support the fact that supplemental agar testing for vancomycin-resistant E. faecalis and E. faecium is not required when the Phoenix instrument is used. However, the Phoenix instrument does have an FDA limitation for the reporting of vancomycin susceptibility for E. casseliflavus and E. gallinarum. The results are suppressed, and the user is directed to test these isolates by another method.
In summary, the Phoenix system compared favorably with the traditional biochemical methods and cell wall fatty acid analysis by GLC for the identification of clinical and challenge strains of 23 species of staphylococci and enterococci. The results of susceptibility testing compared favorably as well. The VME of oxacillin for S. aureus was 1.7% (three isolates). Although this rate is low, it may still warrant supplemental screening for MRSA, particularly for critical isolates such as those from blood cultures. All 36 VREs were reliably identified by the Phoenix system. Although the numbers of other resistance phenotypes such as VISA and enterococci with high-level resistance to aminoglycosides were small in this evaluation, those resistance patterns were likewise accurately detected.
ACKNOWLEDGMENTS
This study was supported in part with a grant from BD Diagnostics, Inc.
We thank the JHH Microbiology Laboratory staff, especially Mark Romagnoli, for their cooperation with this study.
Present address: Waitemata District Health Board, Department Medicine, Level 3, North Shore Hospital, Private Bag, 93-05, Shakespeare Road, Takapuna, Auckland 9, New Zealand.
REFERENCES
Bannerman, T. L. 2003. Staphylococcus, Micrococcus, and other catalase-positive cocci that grow aerobically, p. 384-404. In P. R. Murray, E. J. Baron, J. H. Jorgensen, M. A. Pfaller, and R. H. Yolken (ed.), Manual of clinical microbiology, 8th ed. American Society for Microbiology, Washington, D.C.
Donay, J. L., D. Mathieu, P. Fernandes, C. Pregermain, P. Bruel, A. Wargnier, I. Casin, F. X. Weill, P. H. Lagrange, and J. L. Hermann. 2004. Evaluation of the automated Phoenix system for potential routine use in the clinical microbiology laboratory. J. Clin. Microbiol. 42:1542-1546.
Fahr, A. M., U. Eigner, M. Armburst, A. Caganic, G. Dettori, C. Chezzi, L. Bertoncini, M. Benecchi, and M. G. Menozzi. 2003. Two-center collaborative evaluation of the performance of the BD Phoenix automated microbiology system for identification and susceptibility testing of Enterococcus spp. and Staphylcoccus spp. J. Clin. Microbiol. 41:1135-1142.
Francis, J. S., M. C. Doherty, U. Lopatin, C. P. Johnston, G. Sinha, T. Ross, M. Cai, N. N. Hansel, T. Perl, J. R. Ticehurst, K. Carroll, D. L. Thomas, E. Nuermberger, and J. G. Bartlett. 2005. Severe community-onset pneumonia in healthy adults caused by methicillin-resistant Staphylococcus aureus carrying the Panton-Valentine leukocidin genes. Clin. Infect. Dis. 40:100-107.
Fridkin, S. K., J. C. Hageman, M. Morrison, L. T. Sanza, K. Como-Sabetti, J. A. Jernigan, K. Harriman, L. H. Harrison, R. Lynfield, M. M. Farley, and Active Bacterial Core Surveillance Program of the Emerging Infections Program Network. 2005. Methicillin-resistant Staphylococcus aureus disease in three communities. N. Engl. J. Med. 352:1436-1444.
Horstkotte, M. A., J. K.-M. Knobloch, H. Rohde, S. Dobinsky, and D. Mack. 2004. Evaluation of the BD Phoenix automated microbiology system for detection of methicillin resistance in coagulase-negative staphylococci. J. Clin. Microbiol. 42:5041-5046.
Jones, R. M., M. A. Jackson, C. Ong, and G. K. Lofland. 2002. Endocarditis caused by Staphylococcus lugdunensis. Pediatr. Infect. Dis. J. 21:265-268.
Jorgensen, J. H., and M. J. Ferraro. 2000. Antimicrobial susceptibility testing: special needs for fastidious organisms and difficult-to-detect resistance mechanisms. Clin. Infect. Dis. 30:799-808.
Muto, C. A., J. A. Jernigan, B. E. Ostrowsky, H. M. Richet, W. R. Jarvis, J. M. Boyce, and B. M. Farr. 2003. SHEA guideline for preventing nosocomial transmission of multidrug-resistant strains of Staphylococcus aureus and Enterococcus. Infect. Control Hosp. Epidemiol. 24:362-386.
National Committee for Clinical Laboratory Standards. 2004. Performance standards for antimicrobial susceptibility testing; fourteenth informational supplement. M100-S14. National Committee for Clinical Laboratory Standards, Wayne, Pa.
National Nosocomial Infections Surveillance (NNIS) System. 2004. Data summary from January 1992 through June 2004, issued October 2004. Am. J. Infect. Control 32:470-485.
Robinson, D. A., A. M. Kearns, A. Holmes, D. Morrison, H. Grundmann, G. Edwards, F. G. O'Brien, F. C. Tenover, L. K. McDougal, A. B. Monk, and M. C. Enright. 2005. Re-emergence of early pandemic Staphylococcus aureus as a community acquired methicillin-resistant clone. Lancet 365:1256-1258.
Rupp, M. E., and G. L. Archer. 1994. Coagulase-negative staphylococci: pathogens associated with medical progress. Clin. Infect. Dis. 19:231-243.
Spanu, T., M. Sanguinetti, T. D'Inzeo, D. Ciccaglione, L. Romano, F. Leone, P. Mazzella, and G. Fadda. 2004. Identification of methicillin-resistant isolates of Staphylococcus aureus and coagulase negative staphylococci responsible for bloodstream infections with the PhoenixTM system. Diagn. Microbiol. Infect. Dis. 48:221-227.
Srinivasan, A., J. D. Dick, and T. M. Perl. 2002. Vancomycin resistance in staphylococci. Clin. Microbiol. Rev. 15:430-438.
Stefaniuk, E., E. Baraniak, M. Gniadkowski, and W. Hryniewicz. 2003. Evaluation of the BD Phoenix automated identification and susceptibility testing system in clinical microbiology laboratory practice. Eur. J. Clin. Microbiol. Infect. Dis. 22:479-485.
Steward, C. D., S. A. Stocker, J. M. Swenson, C. M. O'Hara, J. R. Edwards, R. P. Gaynes, J. E. McGowan, Jr., and F. C. Tenover. 1999. Comparison of agar dilution, disk diffusion, MicroScan, and Vitek antimicrobial susceptibility testing methods to broth microdilution for detection of fluoroquinolone-resistant isolates of the family Enterobacteriaceae. J. Clin. Microbiol. 37:544-547.
Tenover, F. C., and L. C. McDonald. 2005. Vancomycin-resistant staphylococci and enterococci: epidemiology and control. Curr. Opin. Infect. Dis. 18:300-305.
Zetola, N., J. S. Francis, E. L. Nuermberger, and W. R. Bishai. 2005. Community acquired methicillin-resistant Staphylococcus aureus: an emerging threat. Lancet Infect. Dis. 5:275-286.(Karen C. Carroll, Anita P)
Microbiology Laboratory, the Johns Hopkins Hospital, Baltimore, Maryland 21287
ABSTRACT
We evaluated the Phoenix automated microbiology system (BD Diagnostic Systems, Sparks, MD) for the identification (ID) and antimicrobial susceptibility testing (AST) of challenge and clinical staphylococci and enterococci recovered from patients in a tertiary-care medical center. In total, 424 isolates were tested: 90 enterococci; 232 Staphylococcus aureus isolates, including 14 vancomycin-intermediate S. aureus isolates; and 102 staphylococci other than S. aureus (non-S. aureus). The Phoenix panels were inoculated according to the manufacturer's instructions. The reference methods for ID comparisons were conventional biochemicals and cell wall fatty acid analysis with the Sherlock microbial identification system (v 3.1; MIDI, Inc. Newark, DE). Agar dilution was the reference AST method. The overall rates of agreement for identification to the genus and the species levels were 99.7% and 99.3%, respectively. All S. aureus isolates and enterococci were correctly identified by the Phoenix panels. For the non-S. aureus staphylococci, there was 98.0% agreement for the ID of 16 different species. The AST results were stratified by organism group. For S. aureus, the categorical agreement (CA) and essential agreement (EA) were 98.2% and 98.8%, respectively. Three of three very major errors (VMEs; 1.7%) were with oxacillin. For non-S. aureus staphylococci, the CA, EA, VME, major errors, and minor error rates were 95.7%, 96.8%, 0.7%, 1.7%, and 2.9%, respectively. The two VMEs were with oxacillin. For the enterococci, there was 100% CA and 99.3% EA. All 36 vancomycin-resistant enterococci were detected by the Phoenix system. The Phoenix system compares favorably to traditional methods for the ID and AST of staphylococci and enterococci.
INTRODUCTION
The burden of the rapid and accurate detection of antimicrobial resistance among gram-positive bacteria remains a continuous challenge for clinical microbiology laboratories. The last decade has seen increases in the numbers of vancomycin-resistant enterococcal (VRE) infections (11, 18), the appearance of glycopeptide resistance among staphylococci (15, 18), a continued rise in nosocomial methicillin-resistant Staphylococcus aureus (MRSA) isolates (9, 11), and a surge in community-associated MRSA isolates (5, 12, 19). In addition, other staphylococci continue to cause serious nosocomial infections, such as endocarditis caused by S. lugdunensis and infections of intravascular and prosthetic devices caused primarily by S. epidermidis (1, 7, 13). This fact, combined with the change in breakpoints for oxacillin/methicillin testing of coagulase-negative staphylococci (10), has presented additional challenges for laboratories, which must adopt systems with improved abilities for the identification (ID) and antimicrobial susceptibility testing (AST) of non-S. aureus staphylococci.
The newer automated instruments have more extensive databases, data management tools (including expert systems), and other features unique to the instrument of each manufacturer. In addition, the accuracy of susceptibility testing has improved with these systems, although problems with the detection of some resistance phenotypes remain (e.g., glycopeptide resistance among S. aureus), leading to recommendations for the continued use of manual screening methods (8, 15, 18).
The newest system to obtain FDA clearance is the Phoenix automated microbiology system (BD Diagnostic Systems, Sparks, MD). This automated system identifies a broad range of gram-positive and gram-negative bacteria. The system includes an instrument, software, disposable panels, broths for ID and AST, and a susceptibility testing indicator. The ID method uses modified conventional, fluorogenic, and chromogenic substrates. The instrument can analyze up to 100 ID and AST combination panels simultaneously. The disposable test panels contain 136 microdilution wells and are available in ID-only, ID/AST, and AST-only formats. The panels are read every 20 min by the instrument.
This study evaluated the performance of the Phoenix instrument for the identification and susceptibility testing of challenge and clinical staphylococci and enterococci isolated from a variety of specimen sources.
MATERIALS AND METHODS
Bacterial strains. One hundred sixty-two previously characterized "challenge" gram-positive cocci and 262 clinical isolates prospectively recovered from routine cultures in the Clinical Microbiology Laboratory from the Johns Hopkins Hospital (JHH) were included in this comparison (the total number of isolates tested was 424). The challenge isolates included 107 S. aureus isolates, including 14 vancomycin-intermediate S. aureus strains (VISA) and 79 clinical S. aureus isolates that were characterized by staphylococcal cassette chromosome mec typing (SCCmec), Panton-Valentine leukocidin (PVL) gene testing, and pulsed-field gel electrophoresis (PFGE) by previously published methods (4). The 79 S. aureus clinical isolates were obtained during an epidemiological study of community-associated MRSA. Also among the challenge isolates were 6 enterococci and 49 coagulase-negative staphylococci that were previously either difficult to identify or that previously had susceptibility patterns that had required supplemental testing. The non-S. aureus challenge isolates included the following: S. epidermidis ATCC 14990, S. saprophyticus ATCC 15305, S. cohnii ATCC 49330, S. warneri ATCC 27836, S. capitis subsp. capitis ATCC 49324, S. chromogenes ATCC 43764, S. intermedius ATCC 29663, S. equorum ATCC 43958, S. gallinarum ATCC 35539, S. kloosii, ATCC 43959, and Macrococcus caseolyticus ATCC 13548.
The prospective clinical isolates included 125 S. aureus isolates, 53 coagulase-negative staphylococci, and 84 enterococcal species. All organisms except for the 14 VISA isolates were from unique patients. Eleven VISA isolates (MICs, 4 to 16 μg/ml) were recovered over time from the same patient who had been treated with prolonged antibiotics for osteomyelitis. In addition, three strains of VISA (MICs, 4 to 8 μg/ml by broth dilution) from wounds or tissues from two sites within the United States and one from Brazil (designated strains NRS 1, NRS 56, and NRS 73, respectively) were obtained through the Network on Antimicrobial Resistance in Staphylococcus aureus (supported under NIAID, NIH, contract no. N01-AI-95359). These VISA isolates are not included in the tables of cumulative data but are addressed separately in the Results section.
Reference identification. The laboratory's routine method for the identification of gram-positive organisms includes testing by the following conventional rapid tests and by assays with biochemicals incorporated into agar media. For S. aureus, slide coagulase and exogenous nuclease tests were routinely performed. If needed, the tube coagulase test, polymyxin B susceptibility testing, a test for ornithine decarboxylation, and a test for fermentation of mannitol were additionally performed. All reactions with the exception of the slide coagulase reaction were read at 24 h. For species-level identification of coagulase-negative staphylococci, a combination of conventional biochemical tests and cell wall fatty acid analysis was performed. The biochemical assays performed included tests for fermentation of sucrose, lactose, mannitol, arabinose, turanose, trehalose, and mannose; a urease detection test; and novobiocin and polymyxin B susceptibility testing. Cellular fatty acid analysis was performed by using the Sherlock microbial identification system (MIDI, Newark, DE). Organism identification was based on computer comparison of the unknown organism's fatty acid methyl ester profile with predetermined fatty acid methyl ester profiles in the library of the microbial identification system with software version 3.1. Sugar fermentation was determined by using a peptone agar base with phenol red indicator. All reactions were interpreted according to published guidelines (1). The enterococci were tested for the following: bile esculin; 6.5% sodium chloride; motility; colony pigmentation; and fermentation of sucrose, lactose, mannitol, sorbitol, arabinose, and sorbose.
Phoenix system identification method. The Phoenix identification method uses modified conventional, fluorogenic, and chromogenic substrates. Combination panels for investigational use only (PMIC/ID-33, catalog no. 448587) for both identification and susceptibility testing were used for this comparison. Software V3.34A/V3.54A was used for this study. The ID side contains 45 wells with dried biochemical substrates and 2 fluorescent control wells. The ID broth was inoculated with bacterial colonies adjusted to a 0.5 McFarland standard by using a CrystalSpec nephelometer (BD Diagnostics), according to the manufacturer's recommendations. The suspension was then poured into the ID side of the Phoenix panel after an aliquot (25 μl) was removed for AST. The specimen was logged and loaded into the instrument within the specified timeline of 30 min. Quality control and maintenance were performed according to the manufacturer's recommendations.
Reference AST. Agar dilution was performed with plates prepared in-house according to CLSI (formerly NCCLS) guidelines (10). The following antibiotics were tested at the indicated concentrations: penicillin, 0.1 to 8 μg/ml; oxacillin, 0.125 to 2 μg/ml; erythromycin, 0.5 to 4 μg/ml; clindamycin, 0.5 to 2 μg/ml; ampicillin, 2 to 16 μg/ml; trimethoprim-sulfamethoxazole, 0.5 and 2 μg/ml; gatifloxacin, 0.5 to 4 μg/ml; vancomycin, 1 to 16 μg/ml; and nitrofurantoin, 32 and 64 μg/ml (urine samples only). For enterococci, a screen for high-level aminoglycoside resistance was also included. MecA gene testing with the LightCycler instrument (Roche Molecular Systems, Pleasanton, CA) was used to confirm any discrepant results for oxacillin obtained by the agar method and with the Phoenix instrument. To determine whether the discrepancies for gatifloxacin were related to methodological differences between an agar-based susceptibility testing method and a broth-based system (as has been reported for other quinolones and members of the family Enterobacteriaceae) (17), broth macrodilution was performed according to CLSI guidelines.
Phoenix system AST. The AST side of the combination panel contains 84 wells with dried antimicrobial panels and one growth control well. The assay is a broth-based microdilution test. The system uses a redox (AST) indicator for the detection of organism growth in the presence of an antimicrobial agent. One free-falling drop of the AST indicator was added to the AST broth tube. Twenty-five microliters of the standardized ID broth suspension was transferred to the AST broth, yielding a final concentration of approximately 5 x 105 CFU/ml. Quality control was performed according to the manufacturer's recommendations. The following antibiotics were tested at the indicated concentrations, and the results were compared to those obtained by agar dilution: penicillin, 0.125 to 16 μg/ml; oxacillin, 0.125 to 4 μg/ml; erythromycin, 0.25 to 4 μg/ml; clindamycin, 0.25 to 4 μg/ml; ampicillin, 0.25 to 32 μg/ml; trimethoprim-sulfamethoxazole, 0.5/9.5 to 2/38 μg/ml; gatifloxacin, 1 to 4 μg/ml; vancomycin, 1 to 16 μg/ml; and nitrofurantoin, 16 to 64 μg/ml (urine samples only). During the study, the version of software that was used had a limitation with the results for coagulase-negative staphylococci (except S. epidermidis) and penicillin. Therefore, fewer comparative results are available for penicillin and coagulase-negative staphylococci. For enterococci, a screen for high-level aminoglycoside resistance for gentamicin and streptomycin was also included. For enterococci, only the results for ampicillin at 0.125 to 32 μg/ml, vancomycin at 1 to 16 μg/ml, tetracycline at 0.5 to 8 μg/ml, and nitrofurantoin at 16 to 64 μg/ml were considered in the data analysis.
Data analysis. All data from both the reference methods and the Phoenix instrument were entered into and analyzed by using Excel spreadsheets. For the ID portion of the test, accuracies to the genus and the species levels were determined. For AST essential agreement (EA) and categorical agreement (CA) were determined. Essential agreement was defined as MICs between the two systems that were within plus or minus 1 doubling dilution. Categorical agreement was defined as susceptible, intermediate, and resistant results that matched between the two systems. The rates of very major errors (VMEs; false-susceptible result with the Phoenix system), major errors (MEs; false-resistant result with the Phoenix system), and minor errors (mEs; one system reported an intermediate result while the other method reported a resistant or susceptible result) were calculated. The number of resistant strains was used as the denominator for the calculation of the VME rate. For the calculation of the ME rates, the number of susceptible strains was used as the denominator.
Discrepant resolution. For identification of discrepancies between the two systems, testing was repeated by both methods. Biochemical testing plus cellular fatty acid analysis by gas-liquid chromatography (GLC) was accepted as the "gold standard." For susceptibility testing, for any organism-drug combination with discrepant results between the two systems, the combination was retested by both methods. Agar dilution testing was accepted as the gold standard for all antimicrobials except oxacillin/methicillin. MecA gene testing was used to resolve the discrepancies obtained by oxacillin/methicillin testing. As stated above, broth macrodilution was used to resolve discrepancies related to gatifloxacin testing.
RESULTS
Table 1 presents the overall results for the identification of the staphylococcal and enterococcal species tested. Initially, 10 isolates (9 non-S. aureus staphylococci and a Macrococcus sp.) had discrepant results between the Phoenix system and conventional methods. Seven of these were resolved by repeat testing with both systems. The overall rates of agreement to the genus and the species levels were 99.7% and 99.3%, respectively, after repeat testing.
As stated above, there were three discrepant results. A Staphylococcus hominis isolate was misidentified by the Phoenix system as a "No ID" initially and then "S. capitis subsp. capitis" on repeat testing, and a Staphylococcus capitis subsp. capitis isolate was misidentified as Staphylococcus hemolyticus. A third discrepant result was obtained for Macrococcus caseolyticus, which was misidentified as S. kloosii. No strains of S. aureus were misidentified as coagulase-negative staphylococci, nor was the opposite true. All Staphylococcus epidermidis isolates and the three isolates of Staphylococcus lugdunensis were correctly identified.
For enterococcal identification, there was 100% agreement to the species level for the 90 strains of E. faecium and E. faecalis. The Phoenix system does not distinguish between E. casseliflavus and E. gallinarum, but it correctly assigned the two organisms in this study to the overlap category. Two E. raffinosus isolates were also correctly identified by the Phoenix instrument.
On initial susceptibility testing, 52 of the 410 clinical and challenge isolates had an organism-drug discrepancy by one of the two systems. After repeat testing, 30 of the isolates gave concordant results by both systems. After repeat testing, the overall EAs for staphylococci and enterococci were 98.2% and 99.3%, respectively. For CA the values were 97.4% and 100% for staphylococci and enterococci, respectively. The overall VME, ME, and mE rates for all organisms were 0.4%, 0.8%, and 1.7%, respectively.
Tables 2 to 4 stratify the susceptibility results by organism group. Table 2 summarizes the results for the 218 S. aureus strains. The overall EA and CA for S. aureus were 98.8% and 98.2%, respectively. Seventy-eight clinical MRSA isolates were recovered during the prospective portion of the study. Within this group, four isolates that tested resistant by the agar dilution method were called susceptible by the Phoenix system. mecA gene PCR confirmed that three of the four isolates were true MRSA isolates. Of these three isolates, the Phoenix system flagged one for additional testing because of its resistance phenotype. However, for two strains that lacked a resistance profile, initial results were susceptible for oxacillin, and the expert system failed to flag these isolates as requiring supplemental testing.
To further challenge the system, a group of 79 diverse MRSA isolates characterized by SCCmec typing, PCR for the Panton-Valentine leukocidin genes, and PFGE were tested. Forty-nine of the isolates were SCCmec type IV, of which 25 were PVL positive; and 23 of the isolates were SCCmec type II, of which none were PVL positive. For seven of the isolates, the SCCmec type could not be determined. The majority of the PVL-positive, SCCmec type IV strains belonged to the USA 300 clone. The Phoenix system called one strain from this group that was a resistant SCCmec type IV strain oxacillin susceptible. When this group was analyzed separately, the EA and CA were 99.6% and 97.8%, respectively (data not shown). When the isolates were stratified by SCCmec type and PFGE patterns, there were no significant differences in the ability of the Phoenix system to accurately identify or perform AST between community-associated MRSA isolates and health care-associated MRSA isolates (data not shown).
Overall, the VME, ME, and mE rates for the total number of S. aureus isolates were 0.4%, 0.3%, and 1.5%, respectively. The three MEs were seen with trimethoprim-sulfamethoxazole. All but four of the mEs were seen with gatifloxacin. Twenty-two strains showed discrepant results for gatifloxacin between the Phoenix system and the agar reference method. Of these 22 strains, 21 of them were intermediate by agar dilution, with an MIC of 4 μg/ml, whereas the Phoenix system called 19 of them susceptible (MIC, 2 μg/ml) and the remaining 2 resistant (MIC, >4 μg/ml). Twenty-one of the strains were retested by broth macrodilution to assess whether the discrepancies were related to methodological differences, as has been reported for other quinolones when other species are tested (17). Fourteen of the 18 (78%) available strains with discrepant intermediate-susceptible results had identical MICs of 2 μg/ml, like the Phoenix results, suggesting that methodological differences exist between agar dilution and broth dilution methods.
Fourteen isolates of MRSA with intermediate resistance to vancomycin by agar dilution (MICs, 4 to 16 μg/ml) were tested. Eleven isolates were recovered over a period of several weeks from cultures of bone and wound specimens from the same JHH patient with MRSA osteomyelitis who had received prolonged vancomycin treatment. The other three isolates were obtained from the Network on Antimicrobial Resistance in Staphylococcus aureus. The Phoenix system MICs for 10 of the JHH strains were 8 μg/ml, and for 1 of the strains the MIC was 16 μg/ml. The MICs of vancomycin for strains NRS 1 and NRS 56 were 8 μg/ml, and for NRS 73 the vancomycin MIC was 4 μg/ml. Based upon these MICs, the Phoenix expert system flagged these isolates as being unusual, suggesting the need for additional tests and, if the finding was verified, notification of infection control personnel. Data for these isolates have been excluded from the cumulative data in Tables 1 and 2.
Table 3 summarizes the findings for the non-S. aureus staphylococci. The overall EA and CA for this group were 96.8% and 95.7%, respectively. The VME, ME, and mE rates were 0.7%, 1.7%, and 2.9%. The two VMEs were seen with oxacillin, whereas trimethoprim-sulfamethoxazole accounted for the majority of the major errors. As was seen with S. aureus, eight of the minor errors were with gatifloxacin.
Table 4 lists the AST results for the 90 enterococcal strains tested. The overall EA and CA for enterococci were 99.3% and 100%, respectively. There were no errors for the enterococci. There were 36 VREs detected by the reference method and with the Phoenix instrument. Fourteen of these isolates were tested for high-level resistance to gentamicin by the reference method, and the result was compared to the Phoenix system's gentamicin high-level resistance result. Of the four isolates with high-level resistance to gentamicin as detected by the reference test, all were also detected by the Phoenix system.
DISCUSSION
The BD Phoenix-100 system is the newest automated instrument to be cleared and marketed by the FDA for use in the United States. This instrument has been available for some years in Europe, and literature on its performance for the identification and susceptibility testing of a broad range of microorganisms is beginning to accumulate. Perhaps of most concern to users is the ability of the system to correctly detect the broad range of resistance phenotypes that have become a universal concern to microbiologists. This study evaluated the reliability of the Phoenix system for the identification and detection of a large number of staphylococci and enterococci representing 23 species and the major resistance phenotypes, including nosocomial and community-associated MRSA clones, VISA, and VRE.
The Phoenix system compared favorably to classical biochemical tests and GLC-based methods for the identification of all of the staphylococci and enterococci tested. The isolates tested represent the strains of coagulase-negative staphylococci and enterococci most commonly recovered in our busy tertiary-care medical center (JHH) but are not inclusive of all possible clinically relevant species. In addition, a limitation of the identification portion of this study is the small numbers of some species available for testing during the evaluation period. The overall concordance rates for the isolates tested were 99.7% to the genus level and 99.3% to the species level. When the results are separated by genus, the Phoenix system had 99.7% and 100% accuracies for the identification of staphylococci and enterococci, respectively. These results are somewhat better than those observed by Donay et al. (2). In that study, there was a 91.8% concordance rate for Staphylococcus sp. identification and an 86.2% concordance rate for the identification of 29 Enterococcus spp. Discrepant results between the Phoenix system and the API 32 Staph identification system (bioMerieux, Inc., Durham, NC) were resolved by 16S rRNA gene sequencing. As in our study, 100% concordance was observed for S. aureus identification. In the two-center study by Fahr et al. (3), 275 staphylococci and 177 enterococci were tested. The rates of concordance were 97.1% and 98.9% for identification of these two groups, respectively, compared to the results obtained with the Vitek 2 and API systems (bioMerieux). None of the discordant results occurred with S. aureus. For the enterococci, two E. faecalis spp. were misidentified as E. casseliflavus-E. gallinarum (3). One hundred percent concordance for S. aureus identification was also seen in the study by Spanu et al. (14), where the overall concordance for the identification of staphylococci to the species level was 98.6% compared to the results obtained with the ID 32 Staph system.
In terms of susceptibility testing, we attempted to challenge the Phoenix with a variety of resistance phenotypes commonly seen in our laboratory environment. Unique to our study, compared to other studies reported in the literature, was the inclusion of a significant number of not only hospital-acquired isolates but also community-associated MRSA strains, as characterized by the presence of PVL genes, SCCmec typing, and PFGE.
The Phoenix system performed well for the detection of MRSA, both community-associated MRSA isolates and health care-associated MRSA isolates. The CA for oxacillin was 98.6%, and the EA for oxacillin was 95.4%. The rates of agreement between the agar dilution method and the Phoenix system were excellent for the other antibiotics evaluated. Fourteen VISA strains were correctly identified by the Phoenix expert system as requiring supplemental testing.
Minor errors were observed primarily for gatifloxacin. No publications have addressed the methodological differences between agar-based methods and broth-based methods for S. aureus when gatifloxacin is tested. Donay et al. (2) tested ofloxacin by agar diffusion and noted a 3.2% mE rate among non-S. aureus staphylococci with the Phoenix system, but this rate was not higher than the rates observed with several other antimicrobial agents. Methodological variations and the high mE rates obtained when members of the family Enterobacteriaceae were tested against two quinolones were reported by Steward et al. (17), who compared a reference broth microdilution method with an agar dilution method. Agar dilution had mE rates of 8.2% for ofloxacin and 12.3% for ciprofloxacin. Gatifloxacin was not tested in that evaluation (17). The limited data obtained by a broth macrodilution method in our study suggest that real methodological differences exist between the agar dilution and the broth dilution methods when staphylococci and gatifloxacin are tested. More studies are needed to confirm our results.
The Phoenix system was able to accurately detect oxacillin resistance among coagulase-negative staphylococci when agar dilution was used as the reference method. There was one VME. In the study by Horstkotte et al. (6), the Phoenix system had a 99.2% sensitivity for the detection of oxacillin/methicillin resistance at the current CLSI (formerly NCCLS) MIC breakpoint of 0.5 μg/ml compared to the results of mecA PCR, which was used as the reference method. By use of this breakpoint, the Phoenix system called 26 mecA-negative strains resistant (specificity, 64.9%). Similar results were obtained by a reference broth microdilution method. The authors concluded that the Phoenix system results were equivalent to those of other phenotypic methods but that confirmation of resistance by mecA PCR should be considered for isolates with oxacillin MICs between 0.5 and 2.0 μg/ml (6).
For enterococcal susceptibility testing, the Phoenix system performed well compared to the performance of the agar dilution method. Vancomycin resistance and -lactam resistance were accurately detected by the instrument. There were no VME or ME for any of the resistance profiles. In addition, high-level resistance to gentamicin was also accurately detected. These results compare favorably to those reported previously by others (2, 3, 16). These data support the fact that supplemental agar testing for vancomycin-resistant E. faecalis and E. faecium is not required when the Phoenix instrument is used. However, the Phoenix instrument does have an FDA limitation for the reporting of vancomycin susceptibility for E. casseliflavus and E. gallinarum. The results are suppressed, and the user is directed to test these isolates by another method.
In summary, the Phoenix system compared favorably with the traditional biochemical methods and cell wall fatty acid analysis by GLC for the identification of clinical and challenge strains of 23 species of staphylococci and enterococci. The results of susceptibility testing compared favorably as well. The VME of oxacillin for S. aureus was 1.7% (three isolates). Although this rate is low, it may still warrant supplemental screening for MRSA, particularly for critical isolates such as those from blood cultures. All 36 VREs were reliably identified by the Phoenix system. Although the numbers of other resistance phenotypes such as VISA and enterococci with high-level resistance to aminoglycosides were small in this evaluation, those resistance patterns were likewise accurately detected.
ACKNOWLEDGMENTS
This study was supported in part with a grant from BD Diagnostics, Inc.
We thank the JHH Microbiology Laboratory staff, especially Mark Romagnoli, for their cooperation with this study.
Present address: Waitemata District Health Board, Department Medicine, Level 3, North Shore Hospital, Private Bag, 93-05, Shakespeare Road, Takapuna, Auckland 9, New Zealand.
REFERENCES
Bannerman, T. L. 2003. Staphylococcus, Micrococcus, and other catalase-positive cocci that grow aerobically, p. 384-404. In P. R. Murray, E. J. Baron, J. H. Jorgensen, M. A. Pfaller, and R. H. Yolken (ed.), Manual of clinical microbiology, 8th ed. American Society for Microbiology, Washington, D.C.
Donay, J. L., D. Mathieu, P. Fernandes, C. Pregermain, P. Bruel, A. Wargnier, I. Casin, F. X. Weill, P. H. Lagrange, and J. L. Hermann. 2004. Evaluation of the automated Phoenix system for potential routine use in the clinical microbiology laboratory. J. Clin. Microbiol. 42:1542-1546.
Fahr, A. M., U. Eigner, M. Armburst, A. Caganic, G. Dettori, C. Chezzi, L. Bertoncini, M. Benecchi, and M. G. Menozzi. 2003. Two-center collaborative evaluation of the performance of the BD Phoenix automated microbiology system for identification and susceptibility testing of Enterococcus spp. and Staphylcoccus spp. J. Clin. Microbiol. 41:1135-1142.
Francis, J. S., M. C. Doherty, U. Lopatin, C. P. Johnston, G. Sinha, T. Ross, M. Cai, N. N. Hansel, T. Perl, J. R. Ticehurst, K. Carroll, D. L. Thomas, E. Nuermberger, and J. G. Bartlett. 2005. Severe community-onset pneumonia in healthy adults caused by methicillin-resistant Staphylococcus aureus carrying the Panton-Valentine leukocidin genes. Clin. Infect. Dis. 40:100-107.
Fridkin, S. K., J. C. Hageman, M. Morrison, L. T. Sanza, K. Como-Sabetti, J. A. Jernigan, K. Harriman, L. H. Harrison, R. Lynfield, M. M. Farley, and Active Bacterial Core Surveillance Program of the Emerging Infections Program Network. 2005. Methicillin-resistant Staphylococcus aureus disease in three communities. N. Engl. J. Med. 352:1436-1444.
Horstkotte, M. A., J. K.-M. Knobloch, H. Rohde, S. Dobinsky, and D. Mack. 2004. Evaluation of the BD Phoenix automated microbiology system for detection of methicillin resistance in coagulase-negative staphylococci. J. Clin. Microbiol. 42:5041-5046.
Jones, R. M., M. A. Jackson, C. Ong, and G. K. Lofland. 2002. Endocarditis caused by Staphylococcus lugdunensis. Pediatr. Infect. Dis. J. 21:265-268.
Jorgensen, J. H., and M. J. Ferraro. 2000. Antimicrobial susceptibility testing: special needs for fastidious organisms and difficult-to-detect resistance mechanisms. Clin. Infect. Dis. 30:799-808.
Muto, C. A., J. A. Jernigan, B. E. Ostrowsky, H. M. Richet, W. R. Jarvis, J. M. Boyce, and B. M. Farr. 2003. SHEA guideline for preventing nosocomial transmission of multidrug-resistant strains of Staphylococcus aureus and Enterococcus. Infect. Control Hosp. Epidemiol. 24:362-386.
National Committee for Clinical Laboratory Standards. 2004. Performance standards for antimicrobial susceptibility testing; fourteenth informational supplement. M100-S14. National Committee for Clinical Laboratory Standards, Wayne, Pa.
National Nosocomial Infections Surveillance (NNIS) System. 2004. Data summary from January 1992 through June 2004, issued October 2004. Am. J. Infect. Control 32:470-485.
Robinson, D. A., A. M. Kearns, A. Holmes, D. Morrison, H. Grundmann, G. Edwards, F. G. O'Brien, F. C. Tenover, L. K. McDougal, A. B. Monk, and M. C. Enright. 2005. Re-emergence of early pandemic Staphylococcus aureus as a community acquired methicillin-resistant clone. Lancet 365:1256-1258.
Rupp, M. E., and G. L. Archer. 1994. Coagulase-negative staphylococci: pathogens associated with medical progress. Clin. Infect. Dis. 19:231-243.
Spanu, T., M. Sanguinetti, T. D'Inzeo, D. Ciccaglione, L. Romano, F. Leone, P. Mazzella, and G. Fadda. 2004. Identification of methicillin-resistant isolates of Staphylococcus aureus and coagulase negative staphylococci responsible for bloodstream infections with the PhoenixTM system. Diagn. Microbiol. Infect. Dis. 48:221-227.
Srinivasan, A., J. D. Dick, and T. M. Perl. 2002. Vancomycin resistance in staphylococci. Clin. Microbiol. Rev. 15:430-438.
Stefaniuk, E., E. Baraniak, M. Gniadkowski, and W. Hryniewicz. 2003. Evaluation of the BD Phoenix automated identification and susceptibility testing system in clinical microbiology laboratory practice. Eur. J. Clin. Microbiol. Infect. Dis. 22:479-485.
Steward, C. D., S. A. Stocker, J. M. Swenson, C. M. O'Hara, J. R. Edwards, R. P. Gaynes, J. E. McGowan, Jr., and F. C. Tenover. 1999. Comparison of agar dilution, disk diffusion, MicroScan, and Vitek antimicrobial susceptibility testing methods to broth microdilution for detection of fluoroquinolone-resistant isolates of the family Enterobacteriaceae. J. Clin. Microbiol. 37:544-547.
Tenover, F. C., and L. C. McDonald. 2005. Vancomycin-resistant staphylococci and enterococci: epidemiology and control. Curr. Opin. Infect. Dis. 18:300-305.
Zetola, N., J. S. Francis, E. L. Nuermberger, and W. R. Bishai. 2005. Community acquired methicillin-resistant Staphylococcus aureus: an emerging threat. Lancet Infect. Dis. 5:275-286.(Karen C. Carroll, Anita P)