Evaluation of the redesigned enterotube--a system for the identification of Enterobacteriaceae

Reevaluation of the API 20E identification system versus conventional biochemicals for identification of members of the family Enterobacteriaceae: a new look at an old product

The API 20E bacterial identification system has been used for 19 years, often as the standard with which other identification systems are compared. Because the accuracy of this system compared with conventional biochemical tests has not been determined in many years, we evaluated the API 20E linear strip by using 291 typical and atypical strains of the family Enterobacteriaceae taken from a culture collection. At 24 h, the API 20E correctly identified by genus and species 229 of 291 (78.7%) of the strains, using Salmonella and Shigella serotyping where indicated. At 48 h, 95.2% were correctly identified by using additional biochemical tests as recommended by the manufacturer. The API 20E misidentified eight (2.7%) strains; these strains were not limited to any particular genus. When 81 of these Enterobacteriaceae strains were arranged into a weighted assortment correlating to the frequency with which they might be found in a clinical laboratory, the API 20E correctly identified 71 (87.7%) at 24 h and 78 (96.3%) at 48 h. This evaluation concluded that the accuracy of the identification of Enterobacteriaceae strains at 24 h (78.7%) may be significantly lower than that of earlier evaluations. However, there is no significant difference in the ability of the API 20E to correctly identify "challenge" type organisms (229 of 291) versus routine hospital isolates (71 of 81) (P greater than 0.05), but the system is not as accurate as the conventional biochemical method of identification.

Commercial identification systems often fail to identify Providencia stuartii

We tested 145 clinical isolates in an attempt to evaluate some of the most widely used commercial identification systems in Europe in terms of their ability to identify Providencia strains. Two manual miniaturized systems (API 20E and Enterotube II) and three mechanized-automated systems (Cobas-Bact, Sceptor System, and Titertek-Enterobac-Rapid Automated System) were evaluated. Providencia alcalifaciens and Providencia rettgeri strains were correctly identified by all systems in all cases, and in most cases identification was achieved without the aid of supplementary tube tests. By contrast, Providencia stuartii was identified without the aid of supplementary tube tests for only 42.5% (API 20E), 37.5% (Enterotube), 68.7% (Sceptor), and 71.2% (Cobas-Bact) of the isolates. The overall misidentification rates were 16.3, 11.3, 11.3, and 10%, respectively. The Titertek-Enterobac-Rapid Automated System failed to identify only 1 of 80 strains (1.3%) and required supplementary tests in 2 other cases (2.5%). Since four of the multitest systems examined often failed to correctly identify P. stuartii, we conclude that supplementary conventional tube tests should always be used to distinguish this species from the other taxa of the Proteeae tribe.

Bacteria on closed-boll and commercially harvested cotton

The bacterial content of specially treated cottons used by other investigators to test human pulmonary responses to cotton dust was examined. Cotton from Lubbock, Tex. and Stoneville, Miss. were either (i) harvested by machine and handled as commercial bale cotton, (ii) harvested as closed bolls with bracts intact and opened under special conditions, (iii) harvested as closed bolls, with bracts being removed and opened under special conditions, or (iv) harvested by (stoneville only). Bacillus spp. were isolated from all samples and predominated in cotton from Stoneville. Enterobacter agglomerans was isolated from all but one sample, the Stoneville closed-boll bract-removed cotton, and predominated in Lubbock samples. Aerogenic and anaerogenic biogroups of E. agglomerans were isolated; only aerogenic strain b of E. agglomerans was present in samples from both locations. Klebsiella ozaenae and K. pneumoniae were isolated only from Lubbock samples. Cotton from Lubbock yielded 100 to 1,000 times more bacteria, both total and gram negative, than did comparably treated cotton from Stoneville. Thus, differences in growing and processing conditions at the two locations were associated with large differences in the bacterial content of the cotton, but harvesting green bolls and removing bracts had little effect. The bacterial content of Stoneville washed cotton, and it paralleled the differences reported (Boehlecke et al., Am. Rev. Respir, Dis. 123:152, 1981) in pulmonary function responses when subjects were exposed to dust (0.6 mg/m3) from these two cottons. Levels of gram-negative and total bacteria on all samples were comparable to those previously reported for field-weathered cottons from various locations throughout the world.

Evaluation of the repliscan system for Enterobacteriaceae identification

A total of 1,013 isolates of Enterobacteriaceae were identified in parallel by the Repliscan (Cathra International, Ontario, Canada) and API 20E (Analytab Products, Plainview, N.Y.) systems. There was a 62% agreement at the genus level between the two systems. Of the 38% discrepant results, Repliscan classified 22% as "biochemical pattern not on file," 8% as a multiple-genus group which included the API 20E identification, and 8% as a genus other than that designated by API 20E. Relative to the various genera, Repliscan agreed with API 20E as follows: Escherichia coli, 80%; Klebsiella spp., 76%; Citrobacter spp., 75%; Proteus spp., 69%; Providencia spp., 54%; Serratia spp., 49%; Enterobacter spp., 25%; Shigella spp., 4%; and Salmonella spp., 0%. Repliscan identified 35% of Enterobacter spp. isolates as Citrobacter spp., 91% of Shigella spp. isolates as a multiple-choice-genus group, and 67% of Salmonella spp. isolates as "biochemical pattern not on file." Repliscan agreed with API 20E at the species level as follows: E. coli, 80%; Klebsiella spp., 56%; Citrobacter spp., 66%; Proteus spp., 55%; Providencia spp., 46%; Serratia spp., 39%; Enterobacter spp., 18%; Shigella spp., 4%; and Salmonella spp., 0%. These findings indicate that the Repliscan system in its present stage of development does not reliably identify the Enterobacteriaceae.

Correlation studies of entero-set 20, API 20E and conventional media systems for Enterobacteriaceae identification

The Entero-Set kit (Fisher Diagnostics) is a 20-biochemical-test system used in the identification of members of the Enterobacteriaceae. This kit was compared with the API 20E (Analytab Products) and conventional media systems, using 505 (303 stock and 202 clinical) strains of Enterobacteriaceae. When the Entero-Set and API 20E results were compared with those of the conventional media system, the Entero-Set performed as well as the API 20E in overall identification. Comparison of common biochemical tests among the various systems showed that citrate, arabinose, adonitol, inositol, and malonate gave correlations below 90%. The majority of the discrepancies were found among stock cultures. In addition, most discrepancies occurred with species of Enterobacter, Salmonella, Proteus, Klebsiella, and Serratia. Reproducibility studies showed the Entero-Set system to perform with a high degree of accuracy and reproducibility.

Evaluation of API 20E System and Encise Enterotube for the identification of Enterobacteriaceae of animal origin

The API 20E System and the Encise Enterotube were evaluated for the identification of the Enterobacteriaceae isolated from clinical specimens of animal origin at a veterinary diagnostic laboratory. Compared to conventional tubed media, the API 20E System identified 235 of 240 isolates (97.9%) correctly. The Encise Enterotube correctly identified 229 of the 240 isolates (95.4%). Thus, both these identification systems could be used to replace conventional methods for identifying members of this family isolated from animal origin.

Bacteria isolated from the duodenum, ileum, and cecum of young chicks

Facultatively anaerobic and strictly anaerobic bacteria colonizing the intestinal tracts of 14-day-old chicks fed a corn-based diet were enumerated, isolated, and identified. Colony counts from anaerobic roll tubes (rumen fluid medium) or aerobic plates (brain heart infusion agar) recovered from homogenates of the duodenum, upper and lower ileum, and cecum varied appreciably among samples from individual birds. Anaerobic and aerobic counts from the duodenum and ileum were similar. Anaerobic counts were highest from the cecum (0.7 X 10(11) to 1.6 X 10(11)/g of dry tissue) and exceeded aerobic plate counts by a factor of at least 10(2). Facultatively anaerobic groups (Streptococcus, Staphylococcus, Lactobacillus, and Escherichia coli) comprised the predominant flora of the duodenum and ileum, although large numbers of anaerobes (9 to 39% of the small intestine isolates), represented by species of Eubacterium, Propionibacterium, Clostridium, Gemmiger, and Fusobacterium, were also recovered. Strict anaerobes (anaerobic gram-positive cocci, Eubacterium, Clostridium Gemmiger, Fusobacterium, and Bacteriodes) made up nearly the entire microbial population of the cecum. Scanning electron microscopy of the intestinal epithelia of chicks revealed populations of microbes on the duodenal, ileal, and cecal mucosal surfaces.

Evaluation of the oxi/ferm tube system with selected Gram-negative bacteria

The Oxi/Ferm test system was evaluated for accuracy and reliability for identification of nonfermentative and oxidase-positive fermentative bacteria by using 375 bacterial strains obtained from stock culture and clinical specimens. The Oxi/Ferm system is a compartmentalized tube containing eight media to provide nine biochemical test results. When combined with the oxidase test, the results corresponding to the positive reactions are totaled and the composite number is located in the coding manual to identify the organisms. The 375 isolates studied were evaluated for accuracy of identification, using both the original and revised code manuals. In comparison with the conventional media used, there was 100% correlation in tests for hydrogen sulfide and indole production, over 96% for nitrogen gas, arginine, and urease, over 92% for xylose and dextrose oxidation, and less than 90% for citrate utilization and dextrose fermentation. There was an overall accuracy in identification of 89.3% using the original manual, with accuracy revised slightly upward to 90.7% using the revised manual. There was 100% accuracy in identification with 44.0% of the strains tested (11 species) using the original manual and with 66.1% (16 species) using the revised manual. Thirteen of the 40 original misidentifications and 14 of 35 revised misidentifications resulted from failure to code and were unidentifiable by Oxi/Ferm. The remainder were incorrectly identified or could not be differentiated from closely related strains. Eleven strains of Alcaligenes odorans were correctly identified using the original code, whereas no code was provided in the revised manual. The Oxi/Ferm system is both simple and rapid and is satisfactory for identification of the more common isolates.

Factors affecting the value of a simple biochemical scheme for identifying Enterobacteriaceae: the reproducible recognition of biotypes

A biochemical screening scheme on agar media for differentiating Enterobacteriaceae in a hospital laboratory is evaluated. Careful observation of test reactions within the scheme permitted the recognition of 78 biotypic reaction patterns which could contribute to epidemiological surveillance. The limitations of the technique are described and discussed and methods of importance in ensuring the reproducibility of reaction patterns emphasised.

Simplified scheme for identification of prompt lactose-fermenting members of the Enterobacteriaceae

A brief, simplified scheme involving the spot indole test and colonial morphology was evaluated for genus level identification of prompt lactose-fermenting (PLF) members of the Enterobacteriaceae. One hundred and ninety-four consecutive, clinically important PLF gram-negative rods isolated in a clinical microbiology laboratory were identified by this simplified scheme, as well as by standard biochemical tests, and the API 20E (Analytab Products, Inc., Plainview, N.Y.) system. In the simplified scheme a flat, spot indole-positive colony was identified as Escherichia coli. Spot indole-negative organisms forming nucoid colonies were identified as Klebsiella sp. or Enterobacter sp. on the basis of semisolid motility and ornithine decarboxylase tests. Approximately 94% of the study isolates followed reactions typical for E. coli, Klebsiella sp., and Enterobacter sp. as defined by this simplified scheme. When compared with the standard and Analytab Products Inc. identifications, the overall accuracy was 97.4%. The accuracy of identification of E. coli, Klebsiella sp., and Enterobacter sp. was 98.1%, 95.6%, and 87.5%, respectively. This simplified scheme is recommended for identification of selected PLF isolates in the clinical microbiology laboratory.

Clinical laboratory evaluation of the further improved Enterotube and Encise II

The most recent improvements in the Enterotube coupled with Encise II permitted identification of routinely and abnormally reacting members of the family Enterobacteriaceae.

Re-evaluation of the Auxotab (Inolex) Enteric 1 System for identification of enterobacteriaceae

The improved Auxotab Enteric I System was re-evaluated and found to correlate 98% with conventional methods for the identification of Enterobacteriaceae. Possible Reasons for previous poor results with this system are discussed.

Rapid methods for biochemical testing of anaerobic bacteria

Rapid biochemical tests for nitrate, indole, gelatin, starch, esculin, and o-nitrophenyl-beta-D-galactopyranoside were performed on 112 strains of anaerobic bacteria. All tests were incubated under aerobic conditions, and results were recorded within 4 h. The tests for nitrate, indole, and starch showed a 95% or greater correlation when compared to the standard biochemical tests. Tests for esculin and gelatin showed an agreement of 86 and 77%, respectively. PathoTec test strips for nitrate, indole, esculin, o-nitrophenyl-beta-D-galactopyranoside, Voges-Proskauer, and urease were also tested and showed encouraging results.