J Clin Pathol 1990;43:587-590 587 Evaluation of Cathra system for identifying Gram

negative aerobic J Clin Pathol: first published as 10.1136/jcp.43.7.587 on 1 July 1990. Downloaded from

J M Ling, L-C Zhang, Y-W Hui, G L French

Abstract Methods The Cathra system is a commercial mul- Five hundred and five recent clinical Gram tipoint inoculation method for the iden- negative bacterial isolates were examined: tification of aerobic Gram negative bac- there were 372 and 133 teria. The system uses a replicator tech- miscellaneous Gram negative bacteria. Five nique in which 21 different agar media reference strains were also tested: Escherichia can be inoculated simultaneously with 36 coli NCTC 10418 and ATCC 25922; organisms. Identifications are made by NCTC 10662 and use of a special computer database. The ATCC 27853; and Salmonella typhimurium performance of this system was com- 42R500. pared with that of the API 20E for the The organisms were identified using the identification of 372 clinical isolates of Cathra System (MCT Medical, Inc, Min- Enterobacteriaceae and 133 miscellan- nesota, USA), and the API 20E (API Sys- eous Gram negative bacteria. For tems, SA, Vercieu, France). The Cathra sys- enterobacteria, the Cathra system was in tem provides 21, and the API 20E 20 bio- 97% -agreement with API 20E at species chemical tests; 13 tests are common to both level and 98% at genus level. For mis- systems (table 1) and indole production must cellaneous Gram negative strains the be tested separately. Tests for bile tolerance, two systems were in 59% agreement at growth on colistin/nalidixic acid, cetrimide/ species level and 77% at genus level. kanamycin and colistin agars, hydrolysis of The Cathra system is suitable for use aesculin, utilisation of malonate, and acidifica- in diagnostic laboratories, especially tion of lactose and cellobiose are available only those with a heavy workload and a wish in the Cathra system; detection of B galac- to use break-point sensitivity testing. tosidase, urea hydrolysis, tryptophan deami- The identification database for mis- nase, Voges-Prokauer reaction, gelatin

cellaneous Gram negative organisms, liquefaction, and acidification of melibiose, http://jcp.bmj.com/ however, needs to be expanded. Table I Biochemical tests provided by Cathra and API 20E identification systemsfor Gram negative aerobic Several commercial systems for the identifica- bacteria tion of Gram negative aerobic organisms are Cathra API 20E available for use in clinical laboratories. Most of these are based on conventional bio- Tolerance to: on September 23, 2021 by guest. Protected copyright. Bile + - chemical tests in dehydrated media which are Colistin/nalidixic acid + - rehydrated by addition of a liquid suspension Cetrimide/kanamycin + - of the test organism. Usually only one or a Colistin + - small number of organisms can be Fermentation of: relatively Mannitol + + tested at one time. The API 20E is the most Lactose + widely used of these systems and is a reference Cellobiose + Glucose + + for the evaluation of other kits.78 The Cathra Arabinose + + System (MCT Medical, Inc, Minnesota, Sorbitol + + Sucrose + + USA) is a new bacterial identification system Rharnnose + + using agar replicator technology. Up to 36 Inositol + + Melibiose - + organisms can be inoculated simultaneously Amygdalin - + in on a variety of identification media standard Possession of: 15 x 10 cm Petri dishes. Twenty one Lysine decarboxylase + + different are available for Ornithine decarboxylase + + Department of agar "repliplates" ,B-Galactosidase - + Microbiology, The the identification of Gram negative non-fas- Tryptophan deaminase - + Chinese University of tidious aerobic bacteria. The manufacturer Hong Kong, The Utilisation of: Prince of Wales can also provide antibiotic plates for sen- Citrate + + Hospital, Shatin, New sitivity testing which can be inoculated at the Malonate + Territories, Hong same time as the biochemical media. Hydrolysis of: Kong Aesculin + J M Ling Results are entered into a "Replianalyzer" Arginine + + L-C Zhang microcomputer which produces printouts of Urea - + Y-W Hui organism identities and suggestions for Production of: G L French additional tests where necessary. Results can Hydrogen sulphide + + Correspondence to: Indole + + also be recorded on a score sheet to generate Acetoin (Voges-Proskauer reaction) - + Dr Julia M Ling an octal code which is used to obtain identities Accepted of publication Liquefaction of gelatin - + 23 January 1990 from a profile register known as "Replidex". 588 Ling, Zhang, Hui, French

Table 2 Comparison of identifications between the The unused tryptone broth cultures were Cathra system and API 20E incubated for a further 14-18 hours at 35°C and the was performed by adding Enterobacteriaceae (n = 372) J Clin Pathol: first published as 10.1136/jcp.43.7.587 on 1 July 1990. Downloaded from Agreement to species level 359 (96-5%) Kovac's reagent.9 to genus level 366 (98-4o/b) In this study the API 20E identities were Miscellaneous Gram negatives (n = 133) assumed to be "correct", and the performance Agreement to species level 79 (59-40) of the Cathra system was assessed against this to genus level 102 (76-7°) standard. amygdalin, and arabinose are provided only in the API 20E system. A single well isolated colony, emulsified in Results 5 ml distilled water, was inoculated into the The number of identities for which the two API 20E system according to the manufac- systems were in agreement is shown in table 2. turer's instructions, and then incubated at For enterobacteria there was 97%o agreement 37°C for 24 hours for enterobacteria and for at sp'ecies level and 98% at genus level. One 48 hours for organisms which did not ferment strain of was identified by glucose and gave less than three positive reac- Cathra as Citrobacterfreundii and one strain of tions. The inoculum for the Cathra system E gergoviae as Klebsiella sp. Among 22 strains was prepared by suspending one to three well of tested, 17 were indole isolated bacterial colonies in 2 ml sterile 1% negative and were identified as P penneri by tryptone broth and incubating until the tur- Cathra. bidity matched that of a MacFarland 0 5 stan- There was less agreement between the two dard. These bacterial suspensions were added systems in the identification of miscellaneous to wells of a seeding tray of which the first was Gram negatives, with 59% agreement at filled with India ink to act as a marker for species level and 7700 at genus level. Thirty orientation. A maximum of 36 different bac- six out of 37 Pseudomonas aeruginosa, both terial cultures could be inoculated with a Plesiomonas shigelloides isolates, and all three replicator device on to the Repliplates. After , however, were correctly iden- the inocula had absorbed into the media the tified by the Cathra system. plates were incubated at 35°C for 16-20 hours. was not in the Cathra database so that the Table 3 Comparison of resultsfor identification ofspecifie Cathra Agreement with API 20E/Number tested "Incorrect" identification*

Enterobacteriaceae http://jcp.bmj.com/ 27/27 24/25 Enterobacter agglomerans 6/6 Enterobacter aerogenes 7/8 E cloacae Enterobacter cloacae 52/53 Citrobacterfreundii Enterobacter agglomerans 2/2 Enterobactergergoviae 0/1 Kiebsiella sp Enterobacter sakazakii 0/5 E cloacae Citrobacterfreundii 39/40 Pseudomonas putida

Citrobacter diversus 16/16 on September 23, 2021 by guest. Protected copyright. Citrobacter amalonaticus 3/3 Edwardsiella tarda 5/5 27/27 Proteus vulgaris 22/22 Morganella morganii 32/33 Proteus mirabilis 26/27 Klebsiella oxytoca 6/8 Aeromonas sp (1) Providencia sp (1) 2/2 Salmonella sp 28/28 Shigella sp 33/33 1/1 Miscellaneous Gram negatives Acinetobacter anitratus 16/23 Citrobacterfreundii (3) Enterobacter sp (2) Aeromonas sp (1) Non-fermenting organism (1) 0/16 Aeromonas sp (16) Pseudomonas aeruginosa 36/37 Xanthomonas maltophilia (1) Pseudomonas cepacia 5/8 Acinetobacter anitratus (2) Serratia sp (1) Pseudomonasfluorescens 1/1 Pseudomonas stutzeri 0/1 Xanthomonas maltophilia (1) Xanthomonas maltophilia 12/17 Pseudomonas cepacia (1) Pseudomonasfluorescens (1) Citrobacterfrewadii (1) Non-fermnting organism (2) Flavobacterium meningosepticum 0/12 Unidentified (10) Xanthomonas maltophilia (1) Aeromonas sp(1) Flavobacterium sp 1/2 Unidentified (1) Plesiomonas shigelloides 2/2 Vibrio cholerae 3/3 3/6 Vibrio sp (1) Salnonellagallinarum (2) 0/1 Aeromonas sp(1) Vibrio vulnificus 0/3 Vibrioparahaemolyticus (3) 0/1 Kkbsiella oxytoca (1) *For the purpose of comparison, API 20E was assumed to be "correct". Cathra identification system 589

three strains were identified as V para- of non-fermenters or Proteus/Providencia sp. haemolyticus. Cathra could identify most Similarly, most enterobacteria would not Acinetobacter anitratus and Xanthomonas mal- grow on cetrimide/kanamycin agar, except for

tophilia, but it was unsuccessful with some Proteus sp and Morganella sp. Scanty J Clin Pathol: first published as 10.1136/jcp.43.7.587 on 1 July 1990. Downloaded from Flavobacterium, and the database did not con- growth on this medium should be regarded as tain F meningosepticum. Neither did the negative. Doubtful results can be entered as database contain species of Aeromonas, so that questionable (?). all 16 A hydrophila strains were identified only The Cathra system is simple and easy to to genus level. use. All the Repliplates are prepared by the manufacturer and sealed in plastic bags to prevent desiccation and contamination during Discussion transport and storage. They are packed in four There have only been two previous sets of five plates each, and each set is clearly evaluations'01' of the Cathra system. Brown labelled together with the date of expiration. and Washington compared the performance of Plates are therefore easily selected and there is Cathra with Enterotube and obtained similar little danger that plates will be missed or results as ours for both Enterobacteriaceae wrongly used. During our study less than 10 and miscellaneous Gram negative organisms.'0 plates out of 300 were found to be contami- Bennett and Joynson, however, found only nated or unusable, despite the subtropical 6500 and 5700 correct enterobacterial iden- climate in Southeast Asia and the need to ship tifications to the genus and species level, res- media from Singapore to Hong Kong. The pectively, when Cathra was compared with plate shelf life is usually more than three API 20E." The relatively poor results weeks and delivery by the manufacturer was obtained in that study may be because they reliable. used the Replidex profile register" while The system is convenient for use in clinical Brown and Washington used the more exten- microbiology laboratories where many organ- sive database of the Replianalyzer computer as isms are isolated each day. Throughput is we did.'0 The low percentages of correct iden- much quicker with this multipoint inoculation tifications for miscellaneous Gram negative system than with identification kits such as organisms in our study was mainly due to API 20E that need to be seeded individually. errors with Flavobacterium sp. The Cathra This methodology is especially convenient for database did not contain F meningosepticum, laboratories that use break-point sensitivity but Cathra could correctly differentiate P vul- testing, and Cathra can also supply the neces- garis and P penneri,2 while both organisms sary antibiotic test plates. were identified as P vulgaris by API 20E. The Replianalyzer microcomputer was a As the formulation of some media in the useful adjunct to the Cathra system. The

Cathra system differs from conventional or programme enables the user to keep records of http://jcp.bmj.com/ the API 20E media, no attempt was made to all the isolates tested and suggests additional compare individual reactions of the systems.1° tests when there is more than one probable No test system can give a "correct" answer for identity for an organism. Supplementary test every organism, but because the API 20E is results can be entered manually into the com- well established and widely accepted,5 we used puter, and the new identity of the organism is it as a standard to evaluate the performance of displayed. Individual result printouts can be Cathra. When disagreement occurred between issued as reports to the wards, while the on September 23, 2021 by guest. Protected copyright. the two systems, tests such as oxidase, laboratory can keep a summary report which motility, and oxidation/fermentation were includes the identifications and biochemical performed as appropriate. These results results. always favoured the identifications provided In summary, the Cathra system is a useful by API 20E. and convenient method for the identification Discrepancies sometimes arose because of of clinical isolates of Enterobacteriaceae. It is difficulties with the interpretation of colour less successful for the identification of the less reactions. Control strains were included in common non-fermenters and the database every test and reproducible results were needs to be extended for this group. The always obtained. Where disagreement multipoint inoculation methodology is very occurred between the two systems for the suitable for busy laboratories and would be identification of test organisms, the strains particularly useful for those using break-point were retested and in most cases the same sensitivity testing. The computed identifica- results were obtained. The Cathra carbo- tion database and report generator provided hydrate acidification plates (lactose, glucose, with the system is helpful and easy to use, but inositol and mannitol), malonate utilisation, inevitably increases the cost of the system. and decarboxylation of ornithine, should be regarded as positive when a very pale pink We are grateful to MCT/Scitech Medical Products, Singa- pore, who supplied us the Repliplates and lent us the Replian- colour is produced. For the detection of H2S alyzer for this study. production, blackening of the colony, with or without diffusion of black pigment into the surrounding medium, should be considered 1 Castillo CB, Bruckner DA. Comparative evaluation of the Eiken and API 20E systems and conventional methods for positive. Scanty growth on colistin/nalidixic identification of members of the family Enterobac- acid agar should be regarded as negative, teriaceae. J Clin Microbiol 1984;20:754-7. 2 Keville MW, Doem GV. Evaluation of the DMS rapid E because Gram negative organisms would not system for identification of clinical isolates of the family grow on this agar except for occasional strains Enterobacteriaceae. J Clin Microbiol 1984;20:1O1O-1 1. 590 Ling, Zhang, Hui, French

3 Fenn JP, Matsen JM. Packaged commercial bacterial iden- 8 Barr JG, Hogg GM, Smyth ETM, Emmerson AM. Com- tification system. Clin Lab Med 1985;5:19-58. parison ofidentification ofEnterobacteriaceae by API 20E 4 Sogaard P, Gahrn-Hansen B, Hui-Ping Z, Frederiksen W. and Sensititre autoidentification system. J Clin Pathol An investigation of three commercial methods for rapid 1989;42:649-52.

identification of non-enteric Gram-negative rods. Acta 9 Cowan ST, Steel KG. Manualfor the identification ofmedical J Clin Pathol: first published as 10.1136/jcp.43.7.587 on 1 July 1990. Downloaded from Pathol Microbiol Immunol Scand (Sect B) 1986;94: bacteria. Cambridge: Cambridge University Press, 357-63. 1985:166. 5 Ling JM, Hui Y-W, French GL. Evaluation of the 10 Brown ST, Washington II JA. Evaluation of the repliscan Microbact-24E bacterial identification system. J Clin system for identification of Enterobacteriaceae. J Clin Pathol 1988;41:910-14. Microbiol 1978;8:695-9. 6 Kampfer P, Dott W. Evaluation of the Titertek-NF system 11 Bennett CHN, Joynson DHM. Kit systems for identifying for identification of Gram-negative nonfermentative and Gram negative aerobic bacilli: report of the Welsh Stand- oxidase-positive fermentative bacteria. J Clin Microbiol ing Specialist Advisory Working Group in Microbiology. 1989;27:1201-5. J Clin Pathol 1986;39:666-71. 7 Appelbaum PC, Arthur RR, Parker ME, Shugar GL, von 12 Farmer III JJ, Davis BR, Hickman-Brenner FW, et al. Kuster LC, Charache P. Comparison ofthree methods for Biochemical identification ofnew species and biogroups of identification of Enterobacteriaceae. Eur J Clin Microbiol Enterobacteriaceae isolated from clinical specimens. J 1982;1:76-81. Clin Microbiol 1985;21:46-76. http://jcp.bmj.com/ on September 23, 2021 by guest. Protected copyright.