methods described by the Virginia Polytechnic Institute and ..... Hazel Livingston, and Theresa Hutchen for assistance in performing the reference identifications.
Vol. 29, No. 3
JOURNAL OF CLINICAL MICROBIOLOGY, Mar. 1991, p. 457-462 0095-1137/91/030457-06$02.00/0 Copyright ©D 1991, American Society for Microbiology
Clinical Evaluation of the RapID-ANA II Panel for Identification of Anaerobic Bacteria DEBORAH M. CELIG1* AND PAUL C. SCHRECKENBERGER" 2 Clinical Microbiology Laboratories, University of Illinois Hospital,' and Department of Pathology,2 University of Illinois at Chicago, Chicago, Illinois 60612 Received 17 October 1990/Accepted 11 December 1990 The accuracy of the RapID-ANA II system (Innovative Diagnostic Systems, Inc., Atlanta, Ga.) was evaluated by comparing the results obtained with that system with results obtained by the methods described by the Virginia Polytechnic Institute and State University. Three hundred anaerobic bacteria were tested, including 259 clinical isolates and 41 stock strains of anaerobic microorganisms representing 16 genera and 48 species. When identifications to the genus level only were included, 96% of the anaerobic gram-negative bacilli, 94% of the Clostridium species, 83% of the anaerobic, nonsporeforming, gram-positive bacilli, and 97% of the anaerobic cocci were correctly identified. When correct identifications to the genus and species levels were compared, 86% of 152 anaerobic gram-negative bacilli, 76% of 34 Clostridium species, 81% of 41 anaerobic, nonsporeforming, gram-positive bacilli, and 97% of 73 anaerobic cocci were correctly identified. Eight isolates (3%) produced inadequate identifications in which the correct identification was listed with one or two other possible choices and extra tests were required for separation. A total of 9 isolates (3%) were misidentified by the RapID-ANA II panel. Overall, the system was able to correctly identify 94% of all the isolates to the genus level and 87% of the isolates to the species level in 4 h by using aerobic incubation.
The purpose of this study was to evaluate the accuracy of the RapID-ANA II for the identification of clinically significant anaerobic bacteria. Results obtained with the RapIDANA II panel were compared with those obtained by the methods described by the Virginia Polytechnic Institute and State University, Blacksburg (13). (Part of this work was presented at the 90th Annual Meeting of the American Society for Microbiology [7].)
Time-consuming biochemical testing under strict anaerobic conditions and gas-liquid chromatographic analysis of short-chain fatty acid metabolites of glucose fermentation have traditionally marked the methods for the identification of anaerobic bacteria (13, 19). These conventional methods are labor intensive, time consuming, expensive, and beyond the means of most clinical microbiology laboratories. Many of the original commercial kit systems were dependent upon the growth of the anaerobic organisms and therefore required 24 to 48 h or more of anaerobic incubation before test results were generated (2). In recent years, kit systems have been developed for the identification of clinically relevant anaerobic bacteria that are not growth dependent and that do not require anaerobic incubation. These systems are based on the detection of preformed bacterial enzymes (23) and include the RapIDANA (Innovative Diagnostic Systems, Inc., Atlanta, Ga.) (1, 3, 4, 6, 8-12, 15, 17, 20, 21, 28), the AN-Ident system (Analytab Products) (4, 6, 12, 21, 27, 29), the Anaerobe Identification card (Vitek Systems, Hazelwood, Mo.) (24), the 2-h ABL system (Austin Biological Systems, Austin, Tex. [22]), the ATB 32A system (API System SA, La Balme les Grottes, Montalieu-Vercieu, France) (18), and the MicroScan system (American MicroScan, Sacramento, Calif.). Innovative Diagnostic Systems, Inc., has revised the original RapID-ANA panel through the deletion of triphenyl tetrazolium reduction, rapid arginine dihydrolase, and trehalose. Three new substrates, namely, urea, p-nitrophenyl-,3-D-disaccharide (BLTS), and p-nitrophenyl-a-L-arabinoside (aARA), have been added to the panel. The RapIDANA II panel and worksheets have been reconfigured for ease of use, and the RapID-ANA II Code Compendium has been revised with updated nomenclature and codes to coincide with the new panel. Evaluations of the newly configured RapID-ANA II panels have been presented recently (5, 26). *
MATERIALS AND METHODS RapID-ANA II panel. The RapID-ANA II panel has 10 reaction wells molded into the periphery of a plastic disposable tray with eight bifunctional wells to allow for a total of 18 biochemical reactions. The panel contains 16 chromogenic substrate tests and 2 modified conventional tests. There are nine nitrophenyl carbohydrate or phosphoester derivatives which, when cleaved by enzymatic hydrolysis, release yellow o- or p-nitrophenol. There are seven P-naphthylamide derivatives of amino acids which, upon enzymatic hydrolysis, release free P-naphthylamine detected by the presence of a purple color after addition of the RapID-ANA II reagent. The RapID-ANA II reagent is a modified cinnamaldehyde reagent for detection of the amine. Modified conventional tests include urea hydrolysis and the formation of indole. The Innova reagent for the detection of indole uses a modification of the spot indole reagent (1% p-dimethylaminocinnamaldehyde with 10% hydrochloric acid). The panel was inoculated by using a pure culture bacterial suspension, equivalent to that of a no. 3 McFarland turbidity standard, prepared in the RapID inoculation fluid from 18 to 72 h of growth on an anaerobic blood plate. The panel was inoculated according to the directions of the manufacturer and incubated aerobically at 35°C for 4 h. After 4 h, the first set of reactions was scored, followed by the addition of the RapID-ANA II reagent and the Innova spot indole reagent to the appropriate wells. A comparator card is available with the kit to facilitate test interpretation. Reactions were re-
Corresponding author. 457
458
CELIG AND SCHRECKENBERGER
corded, and a six-digit profile number was generated. Identifications were obtained by using the RapID-ANA II Code Compendium (16) together with knowledge of the Gram stain reaction, organism morphology, and the profile number. The RapID-ANA II Code Compendium (16) is divided into three major sections: (i) gram-negative anaerobic bacilli, (ii) gram-positive anaerobic bacilli, and (iii) anaerobic cocci. The compendium provides an organism identification or several possible identifications, along with the corresponding probability, biotype, and contraindicated test results. Additional comments associated with the microcode or identification are printed beneath the differential information. Identification to the species level is associated with the following levels of confidence. An "implicit" confidence level indicates that the probability of the first choice is >99.9% and that there are no contraindicated tests. "Satisfactory" indicates a >95.0% probability and only minor contraindicated tests, while an "adequate" confidence level indicates >95.0% probability of the first choice but contraindicated tests resulting in a fair degree of variation from the ideal data base pattern. A "presumptive" identification indicates that there is a single major contraindicated test and the choice should be evaluated to confirm the identification. "Questionable" indicates that the identification is inconsistent and a probability value is therefore not printed. "Inadequate" identifications are generated when probability overlaps occur. "Probability overlap" indicates that more than one choice demonstrates a significant probability of at least 5.0% and may be designated as being among the first two or three choices listed. If the organisms are from the same genus, often a genus- or group-level identification is offered. Taxa contained in the RapID-ANA II Code Compendium (16) include 22 Bacteroides spp., 4 Fusobacterium spp., 3 Porphyromonas spp., Tissierella praeacuta, Wolinella sp., Capnocytophaga sp., and Mobiluncus sp. in the gramnegative rod section. Clostridium clostridiiforme and Clostridium ramosum are included in both the gram-negative and the gram-positive sections of the compendium (16). A total of 24 Clostridium spp., 7 Actinomyces spp., 7 Lactobacillus spp., 3 Eubacterium spp., 2 Propionibacterium spp., Arachnia propionica (Propionibacterium propionicus), Mobiluncus sp., and Bifidobacterium sp. are included in the grampositive section. Eight Peptostreptococcus spp., two
Streptococcus spp., Staphylococcus saccharolyticus, Gemella morbillorum, and Veillonella sp. complete the cocci section of the compendium (16). No attempt is made to differentiate among the species in the genera Capnocytophaga, Mobiluncus, Bifidobacterium, or Veillonella. It has recently been proposed that the moderately saccharolytic Bacteroides spp. that are inhibited by bile be classified in the new genus Prevotella (25). This change has not yet been incorporated into the RapID-ANA II Code Compendium (16), and as such, the genus designation Prevotella is not used in this report. Bacterial strains. The 300 organisms evaluated in the study consisted of 259 clinical isolates and 41 stock strains. Clinical isolates were obtained from patient specimens submitted to the University of Illinois Hospital Anaerobic Bacteriology Laboratory and represented members of the genera Actinomyces, Bacteroides, Bifidobacterium, Capnocytophaga, Clostridium, Eubacterium, Fusobacterium, Gemella, Lacto-
bacillus, Mobiluncus, Peptostreptococcus, Porphyromonas, Propionibacterium, Staphylococcus, Streptococcus, and Veillonella. Stock organisms included organisms from the American Type Culture Collection (Rockville, Md.) and clinical stock strains maintained as quality control organ-
J. CLIN. MICROBIOL.
isms. Stock organisms were subcultured onto agar media at least three times prior to testing with the RapID-ANA II panel and by conventional identification procedures. Conventional identification procedures included Gram staining, aerotolerance, growth on selective and differential media (e.g., kanamycin-vancomycin-laked blood agar, Bacteroides bile esculin agar, egg yolk agar, peptone-yeast extract-glucose broth with bile), and gas-liquid chromatography. Biochemical tests and carbohydrate fermentation reactions were performed with prereduced, anaerobically sterilized media (Carr-Scarborough Microbiologicals, Stone Mountain, Ga.) by methods described in the Virginia Polytechnic Institute's Anaerobe Laboratory Manual (13) and in the latest edition of Bergey's manual (14). Identifications by both conventional tests and with the RapID-ANA II panel were performed in a blinded manner.
RESULTS Table 1 shows the RapID-ANA II panel results for the anaerobic gram-negative bacilli. Of the 65 Bacteroides fragilis group isolates tested, the RapID-ANA II panel correctly identified 53 isolates (81%); all 25 Bacteroides fragilis isolates were identified correctly. Nine B. fragilis group isolates (14%), including seven Bacteroides ovatus and two Bacteroides uniformis isolates, were identified to the genus level only as Bacteroides fragilis group. Two isolates of Bacteroides ovatus were misidentified as Bacteroides thetaiotaomicron, and one strain of Bacteroides uniformis generated an inadequate identification without the correct species included in the identification. When organisms correctly identified to the species or genus level only were included, 62 of 65 B. fragilis group isolates (95%) were correctly identified. Of the 25 Bacteroides melaninogenicus group organisms, 24 (96%) were correctly identified (Table 1). Only one organism was misidentified: a Bacteroides intermedius isolate which was identified by the RapID-ANA II panel as a Bacteroides corporis isolate. Among 47 other Bacteroides spp., 45 isolates (96%) were identified, with 1 strain being misidentified and with 1 strain having an inadequate identification. One Bacteroides oralis isolate was misidentified as Bacteroides oris. An inadequate identification was generated with one isolate of Bacteroides gracilis in which Bacteroides gracilis and Wolinella sp. were listed as the possible identifications. All six Fusobacterium nucleatum isolates were identified to the genus level only, with the use of Gram stain morphology, a lipase reaction, or both given as differential tests. All isolates of Porphyromonas sp., Capnocytophaga sp., and Mobiluncus sp. were correctly identified. One isolate of Bacteroides zoogleoformans and two isolates of Fusobacterium gonidiaformans, which are not included in the RapID-ANA II data base, were called a Bacteroides oralis group isolate and a Fusobacterium sp., respectively. These were the only three organisms tested which were not in the RapID-ANA II data base and, as such, were not included in the statistical evaluation. Of the 152 anaerobic gram-negative bacilli tested, the RapID-ANA II panel identified 131 isolates (86%) to the species level and an additional 15 isolates (10%) to the genus level, with 4 strains (3%) being misidentified and 2 strains (1%) having inadequate identifications. Table 2 presents RapID-ANA II panel results for sporeforming and nonsporeforming, anaerobic, gram-positive bacilli. Of 34 Clostridium spp. tested, 26 (76%) were correctly identified to the species level, including seven Clostridium perfringens, five Clostridium ramosum, 5 Clostridium ca-
VOL. 29, 1991
RapID-ANA II EVALUATION
459
TABLE 1. Identification of anaerobic gram-negative bacilli with the RapID-ANA II panel No. (%) of isolates:
OrganismT
Identified
to
to
Tested
Identified to
species levela
genus levelb
genus or spe-
Bacteroidesfragilis group B. fragilis B. distasonis B. thetaiotaomicron B. ovatus B. vulgatus B. uniformis
65 25 10 10 10 5 5
53 (81.5) 25 (100) 10 (100) 10 (100) 1 (10) 5 (100) 2 (40)
9 (13.8)
2 (3.1)
7 (70)
2 (20)f
Bacteroides melaninogenicus group B. intermedius B. melaninogenicus B. Ioescheii
25 16 8 1
24 (96) 15 (94) 8 (100) 1 (100)
1 (4)f 1 (6Yf
Other Bacteroides spp. B. bivius B. buccae B. capillosus B. oralis B. disiens B. ureolyticus B. gracilis
47 25 11 3 3 2 2 1
45 25 11 3 2 2 2
1 (2/
Inadequate identification, incorrect speciese
1 (1.5)
2 (40)
(96) (100) (100) (100) (67) (100) (100)
1 (20)
1
1 (2)
(33)f 1 (100)
Porphyromonas asaccharolytica
6
Fusobacterium nucleatum
6
Capnocytophaga spp.
2
2 (100)
Mobiluncus spp.
1
1(100)
152
131 (86.2)
Total
Inadequate identification, correct speciesd
6 (100)
6 (100)
15 (9.8)
4 (2.6)
1 (0.7)
1 (0.7)
aImplicit, satisfactory, or acceptable probability of identification to the species level. b Genus- or group-level identification with supplemental tests necessary for resolution of species. Misidentification at the genus level. d Inadequate identification; correct species included among species listed; additional tests necessary. eInadequate identification; correct species not among species listed. f Misidentification at the species level; correct genus identification. "
daveris, 2 Clostridium sordellii, 2 Clostridium septicum, 1 Clostridium clostridiiforme, and 1 Clostridium paraputrificum isolate. Two Clostridium difficile isolates were correctly identified, but one was misidentified as an Eubacterium limosum isolate. Of three Clostridium innocuum isolates tested, one was correctly identified, one yielded an identification of Clostridium spp. in which the correct identification was not among the listed choices, and one was incorrectly identified as a Clostridium subterminale. Of two Clostridium sporogenes isolates tested, both yielded identifications of Clostridium spp., with the choice being between Clostridium difficile and Clostridium sporogenes. All three isolates of Clostridium tertium generated a genus-level identification of Clostridium spp., with the correct choice being among the three possible species and a definitive test of aerobic growth to differentiate the species. Only 2 of 23 isolates of propionibacteria tested gave inadequate identifications in which the correct identification was included in the choices. Both strains were indolenegative strains of Propionibacterium acnes. One isolate of Eubacterium lentum was incorrectly identified as Clostridium hastiforme, one isolate was given an inadequate identification in which the correct identification was among those listed, and five isolates were correctly identified to the
species level. Three isolates of Lactobacillus catenaforme were correctly identified to species level but only one of two isolates of Lactobacillus acidophilus was correctly identified with the second isolate being misidentified as a Lactobacillus catenaforme. Isolates of the genus Bifidobacterium were identified only as Bifidobacterium spp. in the RapID-ANA II system. Two of three isolates were correctly identified. The remaining strain was given an inadequate identification, with the correct identification being among the choices. One isolate of Actinomyces odontolyticus was incorrectly identified as Clostridium hastiforme, while one strain was called Actinomyces sp., with the correct species being among the
possible choices. Table 3 summarizes the RapID-ANA II panel results for the anaerobic cocci. A total of 53 of the peptostreptococci (96%) were correctly identified to the species level. Two Peptostreptococcus prevotii isolates produced inadequate identifications in which the correct identification was among the choices. All isolates of Peptostreptococcus asaccharolyticus, Peptostreptococcus anaerobius, Peptostreptococcus magnus, Peptostreptococcus micros, and Peptostreptococcus tetradius were correctly identified. One Gemella morbillorum, seven Streptococcus intermedius, one Staph-
460
J. CLIN. MICROBIOL.
CELIG AND SCHRECKENBERGER
TABLE 2. Identification of anaerobic sporeforming and nonsporeforming, gram-positive bacilli with the RapID-ANA II panel' No. (%) of isolates:
Identified to
Organism Tested
species level
Identified to
Misidentified
genulev level
species level
genus
to genus or
Clostridium spp. C. perfringens C. ramosum C. cadaveris C. difficile C. innocuum C. tertium C. septicum C. sordellii C. sporogenes C. clostridiiforme C. paraputrificum
34 7 5 5 3 3 3 2 2 2 1 1
26 7 5 5 2 1
Other nonsporeforming spp.
41
33 (81)
Propionibacterium acnes
23
21(91)
Eubacterium lentum
7
5 (71)
1 (14)
Lactobacillus catenaforme Lactobacillus acidophilus
3 2
3 (100) 1 (50)
1 (50)
Bifidobacterium spp.
3
2 (67)
Actinomyces odontolyticus Actinomyces israelii
2 1
1 (100)
75
59 (78.7)
Total
(76) (100) (100) (100) (67) (33)
6 (18)
1 (33) 3 (100)
Inadequate identification, correct species correct species
2 (6)
1 (33) 1 (33)b
2 (100) 2 (100) 2 (100) 1 (100) 1 (100) 1 (2)
3 (7)
4 (10)
2 (9)
1 (14)
1 (33) 1 (50)
1 (50)
7 (9.3)
5 (6.7)
4 (5.3)
a See text and footnotes to Table 1 for explanations of column heads. There were no organisms with inadequate identifications and with the correct species not being among the species listed. b Misidentification at the species level; correct genus identification.
ylococcus saccharolyticus, and nine Veillonella also correctly identified.
spp. were
TABLE 3. Identification of anaerobic gram-positive and gramnegative cocci with the RapID-ANA II panel' No. (%) of isolates:
DISCUSSION In this study, the RapID-ANA II system was evaluated for its ability to identify a wide variety of anaerobic bacteria recovered in the clinical laboratory. After familiarization with the various types of color reactions produced by organisms in the various substrates, the interpretation of most of the tests on the panel was not difficult. The yellow colors resulting from the o- or p-nitrophenol reactions and the dark pink to purple color from the ,3-naphthylamine reactions were generally clear-cut. Both the urea and indole reactions were easily interpreted. In reviewing the initial studies performed on the original RapID-ANA panel, the Anldent panel and the Vitek ANI card, widely varying percentages of accuracy have been reported. In previous studies performed on the first RapIDANA system, 59 to 92% of anaerobic gram-negative bacilli, 66 to 100% of clostridia, 50 to 100% of nonsporeforming gram-positive bacilli, and 78 to 100% of anaerobic cocci were correctly identified by the system compared with correct identifications by various conventional methods (1, 3, 4, 6, 8-12, 15, 17, 20, 21, 28). The wide variations in performance of these systems are related to the numbers, types, and sources (e.g., stock strains, human clinical isolates, and veterinary strains) of organisms tested, whether the additional tests suggested by the systems' computer-generated
Organism
Peptostreptococcus spp. P. asaccharolyticus P. magnus P. anaerobius P. micros P. tetradius P. prevotii
Identified Tested
to.
species level
55 24 13 7 6 3 2
53 (96) 24 (100) 13 (100) 7 (100) 6 (100) 3 (100)
Staphylococcus saccharolyticus
1
1 (100)
Streptococcus intermedius
7
7 (100)
Gemella morbillorum
1
1 (100)
Veillonella sp.
9
9 (100)
73
71 (97.3)
Total
Inadequate identification,
correct species
2 (4)
2 (100)
2 (2.7)
a See text and footnotes to Table 1 for explanations of column heads. There were no organisms identified to only the genus level, misidentified to the genus or species level, or inadequate identification and with the correct species not being among the species listed.
VOL. 29, 1991
codes were included as a part of the "rapid" identification procedure, and whether genus-level identifications were considered as correct identifications. In this evaluation, the RapID-ANA II system correctly identified 131 of 152 anaerobic gram-negative bacilli (86.2%), 26 of 34 Clostridium spp. (76.5%), 33 of 41 nonsporeforming, gram-positive bacilli (80.5%), and 71 of 73 anaerobic cocci (97.3%). When genuslevel identifications are included, 146 of 152 anaerobic gram-negative bacilli (96.1%), 32 of 34 Clostridium spp. (94.1%), 34 of 41 anaerobic, nonsporeforming bacilli (82.9%), and 71 of 73 anaerobic cocci (97.3%) were correctly identified. Seven additional isolates produced inadequate identifications in which the correct identification was among the listed choices and additional instructions were given for further definitive testing. The addition of urea to the RapID-ANA II panel aided in the identification of Bacteroides ureolyticus, Clostridium sordellii, Actinomyces naeslundii, Actinomyces viscosus, and Peptostreptococcus tetradius. The other two new substrates (BLTS and aARA) assisted in differentiating some of the Bacteroides and Clostridium spp. The deletion of trehalose from the RapID-ANA II test panel increased the difficulty of differentiating between Bacteroides ovatus and Bacteroides uniformis. There was no single test in the RapID-ANA II system that separated these two species. Consequently, 7 of 10 Bacteroides ovatus isolates were identified to the group level only as the Bacteroides fragilis group, with Bacteroides ovatus and Bacteroides uniformis being possible identifications of the organism. Two additional strains of Bacteroides ovatus were incorrectly identified as Bacteroides thetaiotaomicron because of positive arginine reactions in the RapID-ANA II panel. The system used arginine and serine as key tests in separating Bacteroides thetaiotaomicron (usually positive) from Bacteroides ovatus and Bacteroides uniformis (usually negative). The RapID-ANA II data base shows that only 11% of Bacteroides ovatus isolates should be positive for arginine. Two strains of Bacteroides uniformis were also identified to the group level only as the Bacteroidesfragilis group, with Bacteroides ovatus and Bacteroides uniformis being the suggested identifications. This was due to a positive reaction with aARA. Bacteroides uniformis is listed as 81% positive for aARA and Bacteroides ovatus is listed as 95% positive for aARA in the RapID-ANA II data base. When this test is negative, the isolate keys out as Bacteroides uniformis. However, when the test is positive, as occurred with two of our isolates, the organism keys out as a Bacteroides fragilis group isolate. One additional strain keyed out as an inadequate identification, with Bacteroides eggerthii and Bacteroides ovatus being the possible choices. The problem here was traced to a negative L-fucoside reaction (Bacteroides uniformis is 95% positive) and a positive glycine reaction (Bacteroides uniformis is 12% positive). Since the indole-positive strains of the Bacteroidesfragilis group are routinely more resistant to a greater number of antibiotics, it is suggested that when reporting the grouplevel identification, the report should be amended to state that the organism is one of the indole-positive strains. Various problems were listed in previously published reports that reviewed the first RapID-ANA system. A major problem cited in several reports addressed the interpretation of the aryl and amide color reactions, variability in test result interpretations by different personnel, and the initial color in the wells becoming obscured by the addition of the secondary reagents (6, 9, 17). The last concern still applies to the RapID-ANA II system; however, interpretation of the color
RapID-ANA II EVALUATION
461
reactions was not considered to be a problem in this study. Referral to the color reaction card available from the manufacturer helped in resolving color interpretations. In this study, all tests were read by one person, eliminating variability in test result interpretations by different personnel. In the study by Karachewski et al. (17), 23% of code numbers generated were not listed in the Code Compendium. In this study, all code numbers were listed in the updated version of the Code Compendium (16). The RapID-ANA II system performed well in identifying certain Clostridium spp., particularly Clostridium perfringens (100%), and Propionibacterium spp. (91%), but it performed less well with other species of gram-positive bacilli and needs to be tested with larger numbers of organisms in order to fully evaluate the system's performance with this group of organisms. The addition of the updated data base, updated nomenclature, new code sheets, and the substitution of three tests has enhanced the system and leads to improved overall performance. In summary, we found the RapID-ANA II panel to be an acceptable rapid test system for identifying most of the clinically significant anaerobic bacteria tested in this study. The decrease in the amount of labor involved in the identification of anaerobic organisms and the more rapid definitive identification help to provide a more timely diagnostic service in clinical anaerobic bacteriology laboratories. ACKNOWLEDGMENTS We thank Gail Richardson, Judy Schermond, Rebecca Shone, Hazel Livingston, and Theresa Hutchen for assistance in performing the reference identifications. REFERENCES 1. Adney, W. S., and R. L. Jones. 1985. Evaluation of the RapIDANA system for identification of anaerobic bacteria of veterinary origin. J. Clin. Microbiol. 22:980-983. 2. Allen, S. D., J. A. Siders, and L. M. Marler. 1985. Isolation and examination of anaerobic bacteria, p. 413-433. In E. H. Lennette, A. Balows, W. J. Hausler, Jr., and H. J. Shadomy (ed.), Manual of clinical microbiology, 4th ed. American Society for Microbiology, Washington, D.C. 3. Appelbaum, P. C., C. S. Kaufmann, and J. W. Depenbusch. 1985. Accuracy and reproducibility of a four-hour method for anaerobe identification. J. Clin. Microbiol. 21:894-898. 4. Bate, G. 1986. Comparison of Minitek Anaerobe II, API An-
Ident, and RapID ANA systems for identification of Clostridium difficile. Am. J. Clin. Pathol. 85:716-718. 5. Beard, M. A., M. L. Scorza, M. Wu, L. C. Gvazdinskas, and W. Landau. 1990. Abstr. Annu. Meet. Am. Soc. Microbiol. 1990, C
51, p. 352. 6. Burlage, R. S., and P. D. Ellner. 1985. Comparison of the PRAS II, AN-Ident, and RapID-ANA systems for identification of anaerobic bacteria. J. Clin. Microbiol. 22:32-35. 7. Celig, D. M., and P. C. Schreckenberger. 1990. Abstr. Annu. Meet. Am. Soc. Microbiol. 1990, C 44, p. 351. 8. Dellinger, C. A., and L. V. H. Moore. 1986. Use of the RapID-ANA System to screen for enzyme activities that differ among species of bile-inhibited Bacteroides. J. Clin. Microbiol. 23:289-293. 9. Downes, J., and J. H. Andrew. 1988. Evaluation of the RapID ANA system as a four-hour method for anaerobe identification. Pathology 20:256-259. 10. Gulletta, E., G. Amato, E. Nani, and I. Covelli. 1985. Comparison of two systems for identification of anaerobic bacteria. Eur. J. Clin. Microbiol. 4:282-285. 11. Harrison, L., M. Dudley, and W. F. Nauschuetz. 1986. Identification of anaerobic bacteria using a rapid system that does not require anaerobic incubation. J. Med. Technol. 3:45-48. 12. Head, C. B., and S. Ratnam. 1988. Comparison of API ZYM
462
13.
14.
15.
16. 17.
18.
19.
CELIG AND SCHRECKENBERGER system with API AN-Ident, API 20A, Minitek Anaerobe II, and RapID-ANA systems for identification of Clostridium difficile. J. Clin. Microbiol. 26:144-146. Holdeman, L. V., E. P. Cato, and W. E. C. Moore (ed.). 1977. Anaerobe laboratory manual, 4th ed. Virginia Polytechnic Institute and State University, Blacksburg. Holdeman, L. V., R. W. Kelley, and W. E. C. Moore. 1984. Family I. Bacteroidaceae Pribram 1933, 10AL, p. 602-662. In N. R. Kreig and J. G. Holt (ed.), Bergey's manual of systematic bacteriology, vol. 1. The Williams & Wilkins Co., Baltimore. Hussain, Z., R. Lannigan, B. C. Schieven, L. Stoakes, T. Kelly, and D. Groves. 1987. Comparison of RapID ANA and Minitek with a conventional method for biochemical identification of anaerobes. Diagn. Microbiol. Infect. Dis. 6:69-72. Innovative Diagnostic Systems, Inc. 1989. RapID ANA II code compendium. Innovative Diagnostic Systems, Inc., Atlanta. Karachewski, N. O., E. L. Busch, and C. L. Wells. 1985. Comparison of PRAS II, RapID ANA, and API 20A systems for identification of anaerobic bacteria. J. Clin. Microbiol. 21:122126. Kitch, T. T., and P. C. Appelbaum. 1989. Accuracy and reproducibility of the 4-hour ATB 32A method for anaerobe identification. J. Clin. Microbiol. 27:2509-2513. Koneman, E. W., S. D. Allen, V. R. Dowell, Jr., W. M. Janda, H. M. Sommers, and W. C. Winn, Jr. 1988. Color atlas and textbook of diagnostic microbiology, 3rd ed. J. B. Lippincott Co., Philadelphia.
20. Murdoch, D. A., I. J. Mitchelmore, R. A. Nash, J. M. Hardie, and S. Tabaqchali. 1988. Preformed enzyme profiles of reference strains of gram-positive anaerobic cocci. J. Med. Microbiol. 27:65-70.
J. CLIN. MICROBIOL. 21. Murray, P. R., C. J. Weber, and A. C. Niles. 1985. Comparative evaluation of three identification systems for anaerobes. J. Clin. Microbiol. 22:52-55. 22. Ortisi, G., M. Cernuschi, P. Scarpellini, A. Pagano, and G. Privitera. 1988. Abstr. Annu. Meet. Am. Soc. Microbiol. 1988, C 298, p. 381. 23. Schreckenberger, P. C., and D. J. Blazevic. 1974. Rapid methods for biochemical testing of anaerobic bacteria. Appl. Microbiol. 28:759-762. 24. Schreckenberger, P. C., D. M. Celig, and W. M. Janda. 1988. Clinical evaluation of the Vitek ANI card for identification of anaerobic bacteria. J. Clin. Microbiol. 26:225-230. 25. Shah, H. N., and D. M. Collins. 1990. Prevotella, a new genus to include Bacteroides melaninogenicus, and related species formerly classified in the genus Bacteroides. Int. J. Syst. Bacteriol. 40:205-208. 26. Siders, J. A., L. C. Wolters, Y. Pettigrew, B. L. Skitt, L. M. Marler, and S. D. Allen. 1990. Abstr. Annu. Meet. Am. Soc. Microbiol. 1990, C 50, p. 352. 27. Stenson, M. J., D. T. Lee, J. E. Rosenblatt, and J. M. Contezac. 1986. Evaluation of the Anident system for the identification of anaerobic bacteria. Diagn. Microbiol. Infect. Dis. 5:9-15. 28. Summanen, P., and H. Jousimies-Somer. 1988. Comparative evaluation of RapID ANA and API 20A for identification of anaerobic bacteria. Eur. J. Clin. Microbiol. Infect. Dis. 7:771775. 29. Tanner, A. C. R., M. N. Strzempko, C. A. Belsky, and G. A. McKinley. 1985. API ZYM and API An-Ident reactions of fastidious oral gram-negative species. J. Clin. Microbiol. 22: 333-335.
