- Email: [email protected]
Alert blood cultures using MicroScan overnight and rapid panels
Direct Bacterial Identification from Positive BacT/Alert Blood Cultures Using MicroScan Overnight and Rapid Panels Ken B. Waites, Eneida S. Brookings, Stephen A. Moser, and Barbara L. Zimmer
Studies were conducted on a method of direct inoculation of MicroScan overnight and rapid panels from positive BacT/ Alert blood culture bottles containing standard aerobic media to determine the correlation with inoculation of the corresponding panels with a standardized bacterial suspension obtained following subculture to agar. For Gram-negative organisms, 122 of 127 (96%) overnight panels and 85 of 118 (72%) rapid panels showed complete agreement with the standard method for species identification. Highest concordance (99%) occurred with Enterobacteriaceae inoculated directly into overnight panels. For Gram-positive organisms, 70 of 85
(82%) overnight panels and 45 of 86 (52%) rapid panels showed complete agreement. These findings suggest that direct inoculation of Gram-negative overnight MicroScan panels yields results most comparable to standard methods when Enterobacteriaceae are detected and allows reporting of results 18 to 24 h sooner. Direct inoculation of Gram-positive overnight or rapid panels and Gram-negative rapid panels from this blood culture medium did not yield acceptable identification results and is not recommended. © 1998 Elsevier Science Inc.
INTRODUCTION
(Chorny and Wilson 1994; Devitt et al. 1995; Doern et al. 1994). Continuously monitoring automated blood culture systems such as BacT/Alert (Organon Teknika Corp., Durham, NC, USA) minimize the time required for detection of bacteremia, frequently indicating a positive culture within 24 h of initial inoculation (Wilson et al. 1994). Conventional technologies require positive bottles to be subcultured to solid media and incubated overnight to provide isolated colonies that are then used to prepare standardized suspensions for speciation of bacterial isolates biochemically, and for determination of antimicrobial susceptibilities (Mirrett 1994). In order to reduce further the turnaround time from initial detection of bacteremia until identification of the etiologic organism, numerous studies have been conducted in which fluid from positive blood cultures is inoculated directly into manual or automated systems used for biochemical speciation of bacteria (Chorny and Wilson 1994; Devitt et al.
Identification of microorganisms causing bacteremia is crucial to guide clinicians in selection of appropriate treatment. Although there is debate about the true value of rapid bacterial identification on patient outcomes, recent studies suggest that availability of such data as soon as possible after bacteremia is confirmed can have a major impact on the care of hospitalized patients with infection in terms of reducing costs for pharmacy, laboratory, other general charges, and may be related to lower mortality From the Department of Pathology, University of Alabama at Birmingham, Birmingham, Alabama, USA (KBW, ESB, SAM), and Dade MicroScan, Inc., West Sacramento, CAUSA (BLZ). Address reprint requests to Ken B. Waites, M.D., Department of Pathology, WP 230, University of Alabama at Birmingham, Birmingham, AL 35233. Presented in part at the 97th General Meeting of the American Society for Microbiology, Miami Beach, Florida, May, 1997 (Abstract C-429). Received 11 February 1998; accepted 24 April 1998.
DIAGN MICROBIOL INFECT DIS 1998;32:21–26 © 1998 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010
0732-8893/98/$19.00 PII S0732-8893(98)00058-3
K.B. Waites et al.
22 1995; Dipersio et al. 1984; Edberg et al. 1975; Hamoudi et al. 1984; Malloy et al. 1983; Moore et al. 1981; Pettigrew et al. 1993). However, there are no published reports regarding the accuracy of bacterial speciation determined by direct inoculation of positive blood cultures detected by the BacT/Alert system into the MicroScan WalkAway System (Dade MicroScan, West Sacramento, CA, USA), using currently available panels, databases, and software. We performed a prospective study in which specially processed fluid from positive BacT/Alert blood culture bottles containing standard aerobic media were inoculated directly into MicroScan overnight and rapid panels. Organism identification was compared with that obtained in standard fashion.
MATERIALS AND METHODS Specimens Tested Aerobic blood cultures collected between September 1995 and April 1996 that were flagged as positive by BacT/Alert and appeared unimicrobic by Gram stain were studied. Specimens that yielded more than one species after subculture were excluded from analysis. Only aerobic and facultative bacteria were included. Hemophilus spp., Neisseria spp., Streptococcus pneumoniae, yeasts, and anaerobes were excluded because the MicroScan panels used are not recommended for identification of these organisms.
MicroScan Instrumentation and Panels The WalkAway/40 with Version 20.30 of the Data Management System was used for all reading and interpretation of results. MicroScan Rapid Neg Combo Type 2 and Dried Overnight Neg Combo Type 15 panels were used for Gram-negative isolates. MicroScan Rapid Pos Combo Type 1 and Overnight Pos Combo Type 6 panels were used for Grampositive isolates.
Blood Culture Instrumentation and Media Blood culture testing was performed on the Organon Teknika BacT/Alert instrument. For clinical purposes, a blood culture set consisted of an aerobic and an anaerobic bottle, each inoculated with 5 to 10 mL of blood. However, direct inoculation studies were limited to positive BacT/Alert aerobic bottles. The standard BacT/Alert blood culture media without any special supplements or antimicrobial inactivating substances was used.
Direct Inoculation Method Fluid (9.5 mL) from aerobic BacT/Alert bottles was inoculated into serum separator tubes (Becton Dick-
inson Vacutainer Systems, Rutherford, NJ, USA) containing 0.2 mL of 1% Triton-X100. Tubes were centrifuged at 1400 3 g for 10 min at room temperature. Bacteria were harvested from the surface of the silicon layer to make an inoculum suspension equivalent to a 0.5% McFarland standard using a MicroScan turbidity meter. MicroScan rapid and overnight panels were then set up and tested following manufacturer’s instructions.
Standard Inoculation Method A small volume of blood culture fluid was inoculated onto trypticase soy agar with 5% sheep blood and MacConkey agar plates with crystal violet and lactose (REMEL, Inc., Lenexa, KS, USA). The blood agar plate was incubated in 5% CO2 and the MacConkey agar plate in air for 18 to 24 h at 35°C to allow bacterial colonies to develop. All panels except rapid Gram-negative panels were inoculated from blood agar plates. MacConkey plates were used to prepare all rapid Gram-negative inocula using standardized suspensions following the manufacturer’s instructions.
Quality Control Quality control organisms recommended by the manufacturer were tested weekly and all results were acceptable. All daily maintenance for the MicroScan WalkAway 40 and the BacT/Alert was performed in accordance with each manufacturer’s instructions.
Data Analysis Direct bacterial identification by rapid panels was compared to standardized testing on rapid panels. Direct identification by overnight panels was compared to standardized testing on overnight panels. Organisms whose probability of identification was ,85%, and those that failed to grow adequately for identification by either overnight or rapid panels inoculated by the standard method, were excluded from analysis. No attempt was made to compare directly the bacterial identification by overnight versus rapid panels inoculated by standard methodology.
RESULTS Two hundred fifty-three (253) blood cultures that appeared unimicrobic on initial Gram stain were evaluated. Sixteen (6.3%) were found to be polymicrobic when subcultured and were excluded from the study. These were predominantly mixtures of
Direct Bacterial Identification
23
Staphylococcus spp. This left 237 potentially evaluable isolates, 133 Gram-negative, and 104 Gram-positive organisms (Table 1). The total numbers of each bacterial species, as identified by standard methodology, using either overnight or rapid panels, are summarized in Table 2. There were some organisms that the WalkAway 40 was unable to adequately identify with overnight and/or rapid panels inoculated by the standard method (Table 1). Rapid panels had more isolates with inadequate growth and/or low probability identifications by the standard method, resulting in fewer evaluable isolates than overnight panels. Among 127 evaluable isolates for the overnight Gram-negative panel, and 118 for the rapid, 96% and 72%, respectively, showed complete agreement between direct and standard methods. The greatest agreement was for Enterobacteriaceae, for which 104 of 105 (99%) isolates showed concordant identification by direct inoculation of overnight panels, versus 77 of 94 (82%) for rapid panels. A single isolate of Escherichia coli identified by the standard method on the Gram-negative overnight panel showed 55% probability by the direct method. Identification of nonfermentative organisms proved more difficult for both standard and direct methods. Twenty-two of 26 (85%) nonfermenters were identified by the standard method using overnight panels, versus 24 of 26 (92%) by rapid panels. Among evaluable nonfermenters, there were 18 of 22 (82%) and 8 of 24 (33%) concordant direct identifications using overnight and rapid panels, respectively.
Table 2 summarizes the discrepant species identifications obtained by direct inoculation in comparison to the standard method. The four Gram-negative nonfermentative organisms with discrepant identities by direct inoculation of overnight panels included two isolates of Acinetobacter spp., one Pseudomonas aeruginosa, and one S. maltophilia. Three that were not identified by direct inoculation were reported as “very rare biotype” (VRB). The fourth, identified as A. baumannii/hemolyticus by standard methods, was called Chryseomonas luteola with 60% probability and A. baumannii/hemolyticus with 15% probability. Gram-positive cocci, predominantly coagulasenegative staphylococci (CNS), also proved difficult to identify to species level by standard methods. There were 85 of 104 (82%) evaluable isolates for overnight Gram-positive panels and 86 of 104 (83%) for rapid panels. There was concordance between direct and standard inoculation for overnight and rapid Gram-positive panels in 70 of 85 (82%) and 45 of 86 (52%) isolates, respectively. Most discrepancies consisted of misclassification among various CNS species, but difficulties distinguishing CNS from S. aureus also occurred, as described in Table 2.
DISCUSSION We evaluated the ability of MicroScan panels inoculated directly from BacT/Alert bottles to identify organisms causing bacteremia in a large university teaching hospital in order to assess whether this
TABLE 1 Bacterial Identification by Direct versus Standard Inoculation of MicroScan Panels Gram-negative Enterobacteriaceae Total unimicrobic isolates tested Overnight panels Evaluable isolates with identification $85% by standard inoculation (%) Evaluable isolates with concordant identification (%) Evaluable isolates with low probability agreement (%) Evaluable isolates with low probability disagreement, complete disagreement, or no identification (%) Rapid panels Evaluable isolates with identification $85% by standard inoculation (%) Evaluable isolates with concordant identification (%) Evaluable isolates with low probability agreement (%) Evaluable isolates with low probability disagreement, complete disagreement, or no identification (%)
Nonfermenters
Total
Gram-positive Total
107
26
133
104
105 (98)
22 (85)
127 (95)
85 (82)
104 (99)
18 (82)
122 (96)
70 (82)
1 (1)
0
1 (0.7)
3 (4)
0
4 (18)
4 (3)
12 (14)
94 (88)
24 (92)
118 (89)
86 (83)
77 (82)
8 (33)
85 (72)
45 (52)
4 (4)
2 (8)
6 (5)
5 (6)
13 (14)
14 (58)
27 (23)
36 (42)
K.B. Waites et al.
24
TABLE 2 Bacteria Identified by Standard Methodology and Summary of Discrepancies by Standard versus Direct Inoculation of MicroScan Panels Overnight Panels
Rapid Panels
Bacteria Identified by Standard Methodology
No. of Isolates
No. of Discrepanciesa
No. of Isolates
No. of Discrepanciesa
Acinetobacter baumannii/haemolyticus Acinetobacter lwoffii Citrobacter freundii Enterobacter aerogenes Enterobacter agglomerans Enterobacter cloacae Escherichia coli Escherichia fergusonii Klebsiella oxytoca Klebsiella pneumoniae Morganella morgannii Ochrobactrum anthropi Proteus mirabilis Providencia rettgeri Pseudomonas aeruginosa Salmonella spp. Serratia marcescens Stenotrophomonas maltophilia Staphylococcus aureus Staphylococcus spp., coagulase-negative Micrococcus spp. Enterococcus faecalis Enterococcus faecium Streptococcus agalactiae Streptococcus equisimilis
3 3 0 3 2 8 47 2 1 30 1 1 7 0 13 1 3 2 35 41 2 6 1 6 2
1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 3b 7c 0 0 0 0 2
5 2 1 4 0 3 47 0 1 28 0 1 3 1 13 3 3 2 25 53 1 6 1 6 2
2 1 1 2 0 1 4 0 1 1 0 1 0 1 9 1 1 1 3d 28e 0 1 1 1 2
a
Includes those isolates with complete discrepancy as well as those discrepant with low probability. One S. aureus was called S. epidermidis, whereas two were called VRB. c All discrepancies involved different species designations of various coagulase-negative Staphylococcus spp. d Three S. aureus were called coagulase-negative Staphylococcus spp., two with low probability. e Six coagulase-negative Staphylococcus spp. were called S. aureus, one with low probability. One S. hyicus, which may be “coagulase variable” was called S. aureus. b
technique could be used to reduce turnaround time and allow appropriate therapy to be applied or verified sooner than if determined by conventional methods. The definitive evaluation of direct bacterial identification would involve a third reference test method for arbitration of discrepancies between the standardized and direct results. This was not attempted because the primary objective was to determine whether direct identification would provide comparable results to those obtained by an automated instrument already in routine use, and because the accuracy of MicroScan panels to identify bacteria when inoculated by standard methodologies has been documented numerous times in comparison to other automated systems and reference methods (Grant et al. 1994; Kloos and George 1991; McGregor et al. 1995; O’Hara and Miller 1992; O’Hara et al. 1993; Pfaller et al. 1991; Rhoads et al. 1995). We found 6.3% of specimens that appeared to be unimicrobic by initial Gram stain were actually polymicrobic on subculture. Using directly inocu-
lated overnight panels, results would be available at the same time as colonial growth on agar from subculture. Thus, polymicrobic specimens could be readily identified before direct inoculation results are reported, thereby reducing the likelihood of misidentification. A more likely occurrence would be that the presence of two or more different organisms would result in low probability or no identification by direct inoculation. This confirmatory approach would not be as practical for directly inoculated rapid panels for which identification would be available up to several hours before detectable growth on agar. Concordant identification in 99% of Enterobacteriaceae by overnight Gram-negative panels inoculated directly suggests this technique may be acceptable for routine use, providing one can be reasonably certain that an organism is not a nonfermenter. This may be possible in many cases because there were no instances in which a nonfermentative organism was misidentified as a member of Enterobacteriaceae, and three out of four discrepant identifications of Gram-
Direct Bacterial Identification negative bacteria in directly inoculated overnight panels yielded reports of VRB. Furthermore, subcultures on blood and MacConkey agars available for examination at the time of the direct identification could be used for further clues to an organism’s identity before reporting results. For example, a Gram-negative bacillus that is oxidase-negative does not resemble Acinetobacter spp. on Gram stain, and if its colonial morphology and characteristics on solid media do not suggest S. maltophilia, it is probably a member of Enterobacteriaceae, and the direct identification is most likely accurate. Although we did not attempt to perform oxidase tests directly on the bacterial pellet before direct inoculation, Dipersio et al. (1984) described a method of direct inoculation of blood cultures in which an oxidase test was performed on a centrifuged pellet of organisms from a positive blood culture bottle to assist in preliminary classification of Gram-negative bacteria. However, Moore et al. (1981) have commented on the difficulty of obtaining a valid oxidase reaction on centrifuged pellets of bacterial cells. Although identification of nonfermenters following direct inoculation appeared to be less reliable than for Enterobacteriaceae, our experience was limited to only 26 isolates. The numbers of nonfermenters identified by standard methods at a low probability also suggest the possible need for an alternative means for routine confirmation. In recent years, CNS have assumed a greater role as nosocomial pathogens, and in some circumstances individual species identification may aid in clinical correlation of disease states and for epidemiologic purposes (Schumacher-Perdreau, 1991). Hence, the ability of automated instruments to distinguish among the many species of CNS can be important and should be assessed. Accurate biochemical speciation of CNS was problematic for panels inoculated by the direct method, as well as those processed by standard methods. A very important and clinically relevant problem with Gram-positive cocci tested with direct inoculation was the inability to distinguish between S. aureus and CNS. This problem was most pronounced with rapid panels. An earlier report (Pettigrew et al. 1993) evaluating direct inoculation of MicroScan rapid panels from positive BacT/Alert bottles detected 89% concordance for direct identification of 83 Gram-negative bacilli versus only 70% for 108 Grampositive cocci. They also noted the inability of directly inoculated rapid panels to distinguish among the various species of CNS.
25 Accuracy of MicroScan panels inoculated directly from positive blood culture bottles for performance of susceptibility testing is another practical consideration that was a separate part of this evaluation because the same panel is typically used for both identification and susceptibility testing, for maximum laboratory efficiency and cost effectiveness. Among the panels evaluated for direct susceptibility testing, only the Gram-negative overnight panel yielded results comparable to standard methods (Waites et al. 1998). With this panel, there was 94.7% overall categorical agreement with the standard method for all drugs tested. The major error (false resistance) rate for directly inoculated panels in comparison to the standard method was 1.4% and very major error (false susceptibility) rate was 2.7%. We conclude that rapid Gram-negative MicroScan panels inoculated directly from positive BacT/Alert bottles do not provide acceptable bacterial identification in comparison to corresponding panels tested by standard methods. In contrast, unimicrobic cultures of Enterobacteriaceae may be appropriately characterized using directly-inoculated Gram-negative overnight panels. Neither rapid nor overnight Grampositive panels accurately identified staphylococci when inoculated directly from positive BacT/Alert bottles and we do not recommend their use for this purpose. After completing this evaluation, we retained our previously existing laboratory policy, identifying bacteria from positive blood cultures using MicroScan overnight panels inoculated following subculture. Laboratories using MicroScan must carefully consider additional costs incurred if direct results for either identification or susceptibility testing must be verified by standard techniques, and the clinical impact of such information for patient care in individual institutional settings. Consideration of alternative less costly nonautomated methods may be more practical if institutional practices support the need for rapid bacterial identification. Laboratories that adapt products such as MicroScan panels for direct bacterial identification from positive blood cultures do so without the benefit of an approved method or quality control. Thus, results must be interpreted and reported with extreme caution.
Financial support for this study was provided by Dade Micro Scan, Inc., West Sacramento, California, and Organon Teknika Corporation, Durham, North Carolina. Technical assistance of Linda Van Pelt and Mary McKinnon is appreciated.
26
K.B. Waites et al.
REFERENCES Chorny JA, Wilson ML (1994) Rapid detection and identification of microorganisms from blood cultures. Clin Lab Med 14:181–196. Devitt KA, O’Sullivan JM, McKinnon ML, Zimmer BL (1995) Feasibility of direct susceptibility testing from various seeded blood culture media into MicroScan rapid and dried overnight susceptibility panels.In: Abstracts of the 95th General Meeting of the American Society for Microbiology. Washington, DC: American Society for Microbiology, abstr. C-345, p. 60. Dipersio JR, Ficorilli SM, Varga FJ (1984) Direct identification and susceptibility testing of gram-negative bacilli from Bactec bottles by use of MS-2 System with updated bacterial identification software. J Clin Microbiol 20: 1202–1204. Doern GV, Vautour R, Gaudet M, Levy B (1994) Clinical impact of rapid in vitro susceptibility and bacterial identification. J Clin Microbiol 32:1757–1762. Edberg SC, Novak M, Slater H, Singer J (1975) Direct inoculation procedure for the rapid classification of bacteria from blood culture. J Clin Microbiol 2:469. Grant CE, Sewell DL, Pflaller MA, Bumgardner RV, Williams JA (1994) Evaluation of two commercial systems for identification of coagulase-negative staphylococci. Diagn Microbiol Infect Dis 18:1–5. Hamoudi AC, Marcon MJ, Cannon HJ (1984) Evaluation of rapid identification of gram-positive cocci in positive blood cultures by use of the AutoMicrobic system grampositive identification card. J Clin Microbiol 20:171–174. Kloos WE, George CG (1991) Identification of staphylococcus species and subspecies with the MicroScan Pos ID and Rapid Pos ID Systems. J Clin Microbiol 29:738–744. Malloy PJ, Ducate MJ, Schreckenberger PC (1983) Comparison of four rapid methods for identification of Enterobacteriaceae from blood cultures. J Clin Microbiol 17:493– 499. McGregor A, Schio F, Beaton S, Boulton V, Perman M, Gilbert G (1995) The MicroScan WalkAway diagnostic microbiology system—an evaluation. Pathology 27:172– 176.
Mirrett S (1994) Antimicrobial susceptibility testing and blood cultures. Clin Lab Med 14:171–179. Moore DF, Hamada SS, Marso E, Martin WJ (1981) Rapid identification and antimicrobial susceptibility testing of gram-negative bacilli from blood cultures by the AutoMicrobic system. J Clin Microbiol 13:934–939. O’Hara CM, Miller JM (1992) Evaluation of the autoScan W/A System for rapid (2 hour) identification of members of the family Enterobacteriaceae. J Clin Microbiol 30:1541–1543. O’Hara CM, Tenover FC, Miller JM (1993) Parallel comparison of accuracy of API 20E, Vitek GNI, MicroScan WalkAway Rapid ID and Becton Dickinson Cobas Micro ID-E/NF for identification of the family Enterobacteriaceae and common gram-negative non-glucose fermenting bacilli. J Clin Microbiol 31:3165–3169. Pettigrew J, Ballou D, Salyer J (1993) Direct susceptibility testing of positive BacT/Alert blood culture systems with autoSCAN-W/A rapid combo panels. In: Abstracts of the General Meeting of the American Society for Microbiology. Washington, DC: American Society for Microbiology, abstr. C-66, p. 547. Pfaller MA, Sahm D, O’Hara C, Ciaglia C, Yu M, Yamane N, Scharnweber G, Rhoden D (1991) Comparison of autoScan W/A rapid bacterial identification system and Vitek AutoMicrobic system for identification of gramnegative bacilli. J Clin Microbiol 29:1422–1428. Rhoads S, Marinelli L, Imperatrice CA, Nachamkin I (1995) Comparison of MicroScan WalkAway System and Vitek System for identification of gram-negative bacteria. J Clin Microbiol 33:3044–3046. Schumacher-Perdreau F (1991) Clinical significance and laboratory diagnosis of coagulase negative staphylococci. Clin Microbiol Newsletter 13:97–101. Waites KB, Brookings ES, Moser SA, Zimmer BL (1998) Direct susceptibility testing from positive BacT/Alert blood cultures by using MicroScan overnight and rapid panels. J Clin Microbiol 36:2052–2056. Wilson, ML, Weinstein, MP, Reller LB (1994) Automated blood culture systems. Clin Lab Med 14:149–169.
Studies were conducted on a method of direct inoculation of MicroScan overnight and rapid panels from positive BacT/ Alert blood culture bottles containing standard aerobic media to determine the correlation with inoculation of the corresponding panels with a standardized bacterial suspension obtained following subculture to agar. For Gram-negative organisms, 122 of 127 (96%) overnight panels and 85 of 118 (72%) rapid panels showed complete agreement with the standard method for species identification. Highest concordance (99%) occurred with Enterobacteriaceae inoculated directly into overnight panels. For Gram-positive organisms, 70 of 85
(82%) overnight panels and 45 of 86 (52%) rapid panels showed complete agreement. These findings suggest that direct inoculation of Gram-negative overnight MicroScan panels yields results most comparable to standard methods when Enterobacteriaceae are detected and allows reporting of results 18 to 24 h sooner. Direct inoculation of Gram-positive overnight or rapid panels and Gram-negative rapid panels from this blood culture medium did not yield acceptable identification results and is not recommended. © 1998 Elsevier Science Inc.
INTRODUCTION
(Chorny and Wilson 1994; Devitt et al. 1995; Doern et al. 1994). Continuously monitoring automated blood culture systems such as BacT/Alert (Organon Teknika Corp., Durham, NC, USA) minimize the time required for detection of bacteremia, frequently indicating a positive culture within 24 h of initial inoculation (Wilson et al. 1994). Conventional technologies require positive bottles to be subcultured to solid media and incubated overnight to provide isolated colonies that are then used to prepare standardized suspensions for speciation of bacterial isolates biochemically, and for determination of antimicrobial susceptibilities (Mirrett 1994). In order to reduce further the turnaround time from initial detection of bacteremia until identification of the etiologic organism, numerous studies have been conducted in which fluid from positive blood cultures is inoculated directly into manual or automated systems used for biochemical speciation of bacteria (Chorny and Wilson 1994; Devitt et al.
Identification of microorganisms causing bacteremia is crucial to guide clinicians in selection of appropriate treatment. Although there is debate about the true value of rapid bacterial identification on patient outcomes, recent studies suggest that availability of such data as soon as possible after bacteremia is confirmed can have a major impact on the care of hospitalized patients with infection in terms of reducing costs for pharmacy, laboratory, other general charges, and may be related to lower mortality From the Department of Pathology, University of Alabama at Birmingham, Birmingham, Alabama, USA (KBW, ESB, SAM), and Dade MicroScan, Inc., West Sacramento, CAUSA (BLZ). Address reprint requests to Ken B. Waites, M.D., Department of Pathology, WP 230, University of Alabama at Birmingham, Birmingham, AL 35233. Presented in part at the 97th General Meeting of the American Society for Microbiology, Miami Beach, Florida, May, 1997 (Abstract C-429). Received 11 February 1998; accepted 24 April 1998.
DIAGN MICROBIOL INFECT DIS 1998;32:21–26 © 1998 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010
0732-8893/98/$19.00 PII S0732-8893(98)00058-3
K.B. Waites et al.
22 1995; Dipersio et al. 1984; Edberg et al. 1975; Hamoudi et al. 1984; Malloy et al. 1983; Moore et al. 1981; Pettigrew et al. 1993). However, there are no published reports regarding the accuracy of bacterial speciation determined by direct inoculation of positive blood cultures detected by the BacT/Alert system into the MicroScan WalkAway System (Dade MicroScan, West Sacramento, CA, USA), using currently available panels, databases, and software. We performed a prospective study in which specially processed fluid from positive BacT/Alert blood culture bottles containing standard aerobic media were inoculated directly into MicroScan overnight and rapid panels. Organism identification was compared with that obtained in standard fashion.
MATERIALS AND METHODS Specimens Tested Aerobic blood cultures collected between September 1995 and April 1996 that were flagged as positive by BacT/Alert and appeared unimicrobic by Gram stain were studied. Specimens that yielded more than one species after subculture were excluded from analysis. Only aerobic and facultative bacteria were included. Hemophilus spp., Neisseria spp., Streptococcus pneumoniae, yeasts, and anaerobes were excluded because the MicroScan panels used are not recommended for identification of these organisms.
MicroScan Instrumentation and Panels The WalkAway/40 with Version 20.30 of the Data Management System was used for all reading and interpretation of results. MicroScan Rapid Neg Combo Type 2 and Dried Overnight Neg Combo Type 15 panels were used for Gram-negative isolates. MicroScan Rapid Pos Combo Type 1 and Overnight Pos Combo Type 6 panels were used for Grampositive isolates.
Blood Culture Instrumentation and Media Blood culture testing was performed on the Organon Teknika BacT/Alert instrument. For clinical purposes, a blood culture set consisted of an aerobic and an anaerobic bottle, each inoculated with 5 to 10 mL of blood. However, direct inoculation studies were limited to positive BacT/Alert aerobic bottles. The standard BacT/Alert blood culture media without any special supplements or antimicrobial inactivating substances was used.
Direct Inoculation Method Fluid (9.5 mL) from aerobic BacT/Alert bottles was inoculated into serum separator tubes (Becton Dick-
inson Vacutainer Systems, Rutherford, NJ, USA) containing 0.2 mL of 1% Triton-X100. Tubes were centrifuged at 1400 3 g for 10 min at room temperature. Bacteria were harvested from the surface of the silicon layer to make an inoculum suspension equivalent to a 0.5% McFarland standard using a MicroScan turbidity meter. MicroScan rapid and overnight panels were then set up and tested following manufacturer’s instructions.
Standard Inoculation Method A small volume of blood culture fluid was inoculated onto trypticase soy agar with 5% sheep blood and MacConkey agar plates with crystal violet and lactose (REMEL, Inc., Lenexa, KS, USA). The blood agar plate was incubated in 5% CO2 and the MacConkey agar plate in air for 18 to 24 h at 35°C to allow bacterial colonies to develop. All panels except rapid Gram-negative panels were inoculated from blood agar plates. MacConkey plates were used to prepare all rapid Gram-negative inocula using standardized suspensions following the manufacturer’s instructions.
Quality Control Quality control organisms recommended by the manufacturer were tested weekly and all results were acceptable. All daily maintenance for the MicroScan WalkAway 40 and the BacT/Alert was performed in accordance with each manufacturer’s instructions.
Data Analysis Direct bacterial identification by rapid panels was compared to standardized testing on rapid panels. Direct identification by overnight panels was compared to standardized testing on overnight panels. Organisms whose probability of identification was ,85%, and those that failed to grow adequately for identification by either overnight or rapid panels inoculated by the standard method, were excluded from analysis. No attempt was made to compare directly the bacterial identification by overnight versus rapid panels inoculated by standard methodology.
RESULTS Two hundred fifty-three (253) blood cultures that appeared unimicrobic on initial Gram stain were evaluated. Sixteen (6.3%) were found to be polymicrobic when subcultured and were excluded from the study. These were predominantly mixtures of
Direct Bacterial Identification
23
Staphylococcus spp. This left 237 potentially evaluable isolates, 133 Gram-negative, and 104 Gram-positive organisms (Table 1). The total numbers of each bacterial species, as identified by standard methodology, using either overnight or rapid panels, are summarized in Table 2. There were some organisms that the WalkAway 40 was unable to adequately identify with overnight and/or rapid panels inoculated by the standard method (Table 1). Rapid panels had more isolates with inadequate growth and/or low probability identifications by the standard method, resulting in fewer evaluable isolates than overnight panels. Among 127 evaluable isolates for the overnight Gram-negative panel, and 118 for the rapid, 96% and 72%, respectively, showed complete agreement between direct and standard methods. The greatest agreement was for Enterobacteriaceae, for which 104 of 105 (99%) isolates showed concordant identification by direct inoculation of overnight panels, versus 77 of 94 (82%) for rapid panels. A single isolate of Escherichia coli identified by the standard method on the Gram-negative overnight panel showed 55% probability by the direct method. Identification of nonfermentative organisms proved more difficult for both standard and direct methods. Twenty-two of 26 (85%) nonfermenters were identified by the standard method using overnight panels, versus 24 of 26 (92%) by rapid panels. Among evaluable nonfermenters, there were 18 of 22 (82%) and 8 of 24 (33%) concordant direct identifications using overnight and rapid panels, respectively.
Table 2 summarizes the discrepant species identifications obtained by direct inoculation in comparison to the standard method. The four Gram-negative nonfermentative organisms with discrepant identities by direct inoculation of overnight panels included two isolates of Acinetobacter spp., one Pseudomonas aeruginosa, and one S. maltophilia. Three that were not identified by direct inoculation were reported as “very rare biotype” (VRB). The fourth, identified as A. baumannii/hemolyticus by standard methods, was called Chryseomonas luteola with 60% probability and A. baumannii/hemolyticus with 15% probability. Gram-positive cocci, predominantly coagulasenegative staphylococci (CNS), also proved difficult to identify to species level by standard methods. There were 85 of 104 (82%) evaluable isolates for overnight Gram-positive panels and 86 of 104 (83%) for rapid panels. There was concordance between direct and standard inoculation for overnight and rapid Gram-positive panels in 70 of 85 (82%) and 45 of 86 (52%) isolates, respectively. Most discrepancies consisted of misclassification among various CNS species, but difficulties distinguishing CNS from S. aureus also occurred, as described in Table 2.
DISCUSSION We evaluated the ability of MicroScan panels inoculated directly from BacT/Alert bottles to identify organisms causing bacteremia in a large university teaching hospital in order to assess whether this
TABLE 1 Bacterial Identification by Direct versus Standard Inoculation of MicroScan Panels Gram-negative Enterobacteriaceae Total unimicrobic isolates tested Overnight panels Evaluable isolates with identification $85% by standard inoculation (%) Evaluable isolates with concordant identification (%) Evaluable isolates with low probability agreement (%) Evaluable isolates with low probability disagreement, complete disagreement, or no identification (%) Rapid panels Evaluable isolates with identification $85% by standard inoculation (%) Evaluable isolates with concordant identification (%) Evaluable isolates with low probability agreement (%) Evaluable isolates with low probability disagreement, complete disagreement, or no identification (%)
Nonfermenters
Total
Gram-positive Total
107
26
133
104
105 (98)
22 (85)
127 (95)
85 (82)
104 (99)
18 (82)
122 (96)
70 (82)
1 (1)
0
1 (0.7)
3 (4)
0
4 (18)
4 (3)
12 (14)
94 (88)
24 (92)
118 (89)
86 (83)
77 (82)
8 (33)
85 (72)
45 (52)
4 (4)
2 (8)
6 (5)
5 (6)
13 (14)
14 (58)
27 (23)
36 (42)
K.B. Waites et al.
24
TABLE 2 Bacteria Identified by Standard Methodology and Summary of Discrepancies by Standard versus Direct Inoculation of MicroScan Panels Overnight Panels
Rapid Panels
Bacteria Identified by Standard Methodology
No. of Isolates
No. of Discrepanciesa
No. of Isolates
No. of Discrepanciesa
Acinetobacter baumannii/haemolyticus Acinetobacter lwoffii Citrobacter freundii Enterobacter aerogenes Enterobacter agglomerans Enterobacter cloacae Escherichia coli Escherichia fergusonii Klebsiella oxytoca Klebsiella pneumoniae Morganella morgannii Ochrobactrum anthropi Proteus mirabilis Providencia rettgeri Pseudomonas aeruginosa Salmonella spp. Serratia marcescens Stenotrophomonas maltophilia Staphylococcus aureus Staphylococcus spp., coagulase-negative Micrococcus spp. Enterococcus faecalis Enterococcus faecium Streptococcus agalactiae Streptococcus equisimilis
3 3 0 3 2 8 47 2 1 30 1 1 7 0 13 1 3 2 35 41 2 6 1 6 2
1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 3b 7c 0 0 0 0 2
5 2 1 4 0 3 47 0 1 28 0 1 3 1 13 3 3 2 25 53 1 6 1 6 2
2 1 1 2 0 1 4 0 1 1 0 1 0 1 9 1 1 1 3d 28e 0 1 1 1 2
a
Includes those isolates with complete discrepancy as well as those discrepant with low probability. One S. aureus was called S. epidermidis, whereas two were called VRB. c All discrepancies involved different species designations of various coagulase-negative Staphylococcus spp. d Three S. aureus were called coagulase-negative Staphylococcus spp., two with low probability. e Six coagulase-negative Staphylococcus spp. were called S. aureus, one with low probability. One S. hyicus, which may be “coagulase variable” was called S. aureus. b
technique could be used to reduce turnaround time and allow appropriate therapy to be applied or verified sooner than if determined by conventional methods. The definitive evaluation of direct bacterial identification would involve a third reference test method for arbitration of discrepancies between the standardized and direct results. This was not attempted because the primary objective was to determine whether direct identification would provide comparable results to those obtained by an automated instrument already in routine use, and because the accuracy of MicroScan panels to identify bacteria when inoculated by standard methodologies has been documented numerous times in comparison to other automated systems and reference methods (Grant et al. 1994; Kloos and George 1991; McGregor et al. 1995; O’Hara and Miller 1992; O’Hara et al. 1993; Pfaller et al. 1991; Rhoads et al. 1995). We found 6.3% of specimens that appeared to be unimicrobic by initial Gram stain were actually polymicrobic on subculture. Using directly inocu-
lated overnight panels, results would be available at the same time as colonial growth on agar from subculture. Thus, polymicrobic specimens could be readily identified before direct inoculation results are reported, thereby reducing the likelihood of misidentification. A more likely occurrence would be that the presence of two or more different organisms would result in low probability or no identification by direct inoculation. This confirmatory approach would not be as practical for directly inoculated rapid panels for which identification would be available up to several hours before detectable growth on agar. Concordant identification in 99% of Enterobacteriaceae by overnight Gram-negative panels inoculated directly suggests this technique may be acceptable for routine use, providing one can be reasonably certain that an organism is not a nonfermenter. This may be possible in many cases because there were no instances in which a nonfermentative organism was misidentified as a member of Enterobacteriaceae, and three out of four discrepant identifications of Gram-
Direct Bacterial Identification negative bacteria in directly inoculated overnight panels yielded reports of VRB. Furthermore, subcultures on blood and MacConkey agars available for examination at the time of the direct identification could be used for further clues to an organism’s identity before reporting results. For example, a Gram-negative bacillus that is oxidase-negative does not resemble Acinetobacter spp. on Gram stain, and if its colonial morphology and characteristics on solid media do not suggest S. maltophilia, it is probably a member of Enterobacteriaceae, and the direct identification is most likely accurate. Although we did not attempt to perform oxidase tests directly on the bacterial pellet before direct inoculation, Dipersio et al. (1984) described a method of direct inoculation of blood cultures in which an oxidase test was performed on a centrifuged pellet of organisms from a positive blood culture bottle to assist in preliminary classification of Gram-negative bacteria. However, Moore et al. (1981) have commented on the difficulty of obtaining a valid oxidase reaction on centrifuged pellets of bacterial cells. Although identification of nonfermenters following direct inoculation appeared to be less reliable than for Enterobacteriaceae, our experience was limited to only 26 isolates. The numbers of nonfermenters identified by standard methods at a low probability also suggest the possible need for an alternative means for routine confirmation. In recent years, CNS have assumed a greater role as nosocomial pathogens, and in some circumstances individual species identification may aid in clinical correlation of disease states and for epidemiologic purposes (Schumacher-Perdreau, 1991). Hence, the ability of automated instruments to distinguish among the many species of CNS can be important and should be assessed. Accurate biochemical speciation of CNS was problematic for panels inoculated by the direct method, as well as those processed by standard methods. A very important and clinically relevant problem with Gram-positive cocci tested with direct inoculation was the inability to distinguish between S. aureus and CNS. This problem was most pronounced with rapid panels. An earlier report (Pettigrew et al. 1993) evaluating direct inoculation of MicroScan rapid panels from positive BacT/Alert bottles detected 89% concordance for direct identification of 83 Gram-negative bacilli versus only 70% for 108 Grampositive cocci. They also noted the inability of directly inoculated rapid panels to distinguish among the various species of CNS.
25 Accuracy of MicroScan panels inoculated directly from positive blood culture bottles for performance of susceptibility testing is another practical consideration that was a separate part of this evaluation because the same panel is typically used for both identification and susceptibility testing, for maximum laboratory efficiency and cost effectiveness. Among the panels evaluated for direct susceptibility testing, only the Gram-negative overnight panel yielded results comparable to standard methods (Waites et al. 1998). With this panel, there was 94.7% overall categorical agreement with the standard method for all drugs tested. The major error (false resistance) rate for directly inoculated panels in comparison to the standard method was 1.4% and very major error (false susceptibility) rate was 2.7%. We conclude that rapid Gram-negative MicroScan panels inoculated directly from positive BacT/Alert bottles do not provide acceptable bacterial identification in comparison to corresponding panels tested by standard methods. In contrast, unimicrobic cultures of Enterobacteriaceae may be appropriately characterized using directly-inoculated Gram-negative overnight panels. Neither rapid nor overnight Grampositive panels accurately identified staphylococci when inoculated directly from positive BacT/Alert bottles and we do not recommend their use for this purpose. After completing this evaluation, we retained our previously existing laboratory policy, identifying bacteria from positive blood cultures using MicroScan overnight panels inoculated following subculture. Laboratories using MicroScan must carefully consider additional costs incurred if direct results for either identification or susceptibility testing must be verified by standard techniques, and the clinical impact of such information for patient care in individual institutional settings. Consideration of alternative less costly nonautomated methods may be more practical if institutional practices support the need for rapid bacterial identification. Laboratories that adapt products such as MicroScan panels for direct bacterial identification from positive blood cultures do so without the benefit of an approved method or quality control. Thus, results must be interpreted and reported with extreme caution.
Financial support for this study was provided by Dade Micro Scan, Inc., West Sacramento, California, and Organon Teknika Corporation, Durham, North Carolina. Technical assistance of Linda Van Pelt and Mary McKinnon is appreciated.
26
K.B. Waites et al.
REFERENCES Chorny JA, Wilson ML (1994) Rapid detection and identification of microorganisms from blood cultures. Clin Lab Med 14:181–196. Devitt KA, O’Sullivan JM, McKinnon ML, Zimmer BL (1995) Feasibility of direct susceptibility testing from various seeded blood culture media into MicroScan rapid and dried overnight susceptibility panels.In: Abstracts of the 95th General Meeting of the American Society for Microbiology. Washington, DC: American Society for Microbiology, abstr. C-345, p. 60. Dipersio JR, Ficorilli SM, Varga FJ (1984) Direct identification and susceptibility testing of gram-negative bacilli from Bactec bottles by use of MS-2 System with updated bacterial identification software. J Clin Microbiol 20: 1202–1204. Doern GV, Vautour R, Gaudet M, Levy B (1994) Clinical impact of rapid in vitro susceptibility and bacterial identification. J Clin Microbiol 32:1757–1762. Edberg SC, Novak M, Slater H, Singer J (1975) Direct inoculation procedure for the rapid classification of bacteria from blood culture. J Clin Microbiol 2:469. Grant CE, Sewell DL, Pflaller MA, Bumgardner RV, Williams JA (1994) Evaluation of two commercial systems for identification of coagulase-negative staphylococci. Diagn Microbiol Infect Dis 18:1–5. Hamoudi AC, Marcon MJ, Cannon HJ (1984) Evaluation of rapid identification of gram-positive cocci in positive blood cultures by use of the AutoMicrobic system grampositive identification card. J Clin Microbiol 20:171–174. Kloos WE, George CG (1991) Identification of staphylococcus species and subspecies with the MicroScan Pos ID and Rapid Pos ID Systems. J Clin Microbiol 29:738–744. Malloy PJ, Ducate MJ, Schreckenberger PC (1983) Comparison of four rapid methods for identification of Enterobacteriaceae from blood cultures. J Clin Microbiol 17:493– 499. McGregor A, Schio F, Beaton S, Boulton V, Perman M, Gilbert G (1995) The MicroScan WalkAway diagnostic microbiology system—an evaluation. Pathology 27:172– 176.
Mirrett S (1994) Antimicrobial susceptibility testing and blood cultures. Clin Lab Med 14:171–179. Moore DF, Hamada SS, Marso E, Martin WJ (1981) Rapid identification and antimicrobial susceptibility testing of gram-negative bacilli from blood cultures by the AutoMicrobic system. J Clin Microbiol 13:934–939. O’Hara CM, Miller JM (1992) Evaluation of the autoScan W/A System for rapid (2 hour) identification of members of the family Enterobacteriaceae. J Clin Microbiol 30:1541–1543. O’Hara CM, Tenover FC, Miller JM (1993) Parallel comparison of accuracy of API 20E, Vitek GNI, MicroScan WalkAway Rapid ID and Becton Dickinson Cobas Micro ID-E/NF for identification of the family Enterobacteriaceae and common gram-negative non-glucose fermenting bacilli. J Clin Microbiol 31:3165–3169. Pettigrew J, Ballou D, Salyer J (1993) Direct susceptibility testing of positive BacT/Alert blood culture systems with autoSCAN-W/A rapid combo panels. In: Abstracts of the General Meeting of the American Society for Microbiology. Washington, DC: American Society for Microbiology, abstr. C-66, p. 547. Pfaller MA, Sahm D, O’Hara C, Ciaglia C, Yu M, Yamane N, Scharnweber G, Rhoden D (1991) Comparison of autoScan W/A rapid bacterial identification system and Vitek AutoMicrobic system for identification of gramnegative bacilli. J Clin Microbiol 29:1422–1428. Rhoads S, Marinelli L, Imperatrice CA, Nachamkin I (1995) Comparison of MicroScan WalkAway System and Vitek System for identification of gram-negative bacteria. J Clin Microbiol 33:3044–3046. Schumacher-Perdreau F (1991) Clinical significance and laboratory diagnosis of coagulase negative staphylococci. Clin Microbiol Newsletter 13:97–101. Waites KB, Brookings ES, Moser SA, Zimmer BL (1998) Direct susceptibility testing from positive BacT/Alert blood cultures by using MicroScan overnight and rapid panels. J Clin Microbiol 36:2052–2056. Wilson, ML, Weinstein, MP, Reller LB (1994) Automated blood culture systems. Clin Lab Med 14:149–169.