Evaluation of the AutoMicrobic system for identification and susceptibility testing of gram-negative bacilli

The AutoMicrobic system (AMS) (Vitek Systems, Inc., Hazelwood, Mo.) was compared with the API-20E system for the identification of gram-negative bacilli by using 380 stock clinical isolates and 377 immediately encountered fresh clinical isolates. For the stock isolates, with Enterobacteriaceae-Plus Biochemical Cards and automated interpretation, 364 (95.8%) were in agreement to the species level. For the fresh clinical isolates, agreement at the genus and species levels was 89.7 and 85.9%, respectively, when Enterobacteriaceae-Plus Cards were interpreted by the AMS. Manual interpretation of Enterobacteriaceae-Plus Biochemical Cards improved species level agreement to 91.0%. Subsequent retesting of all discrepant isolates with the Gram-Negative Identification Card resulted in significant improvement of results, and for the stock and fresh clinical isolates, species level agreement was 98.7 and 97.3%, respectively. AMS susceptibility testing was evaluated by comparing ampicillin and cephalothin MICs determined in parallel by AMS and a reference broth microdilution test for stock isolates, and by comparison of AMS and standardized disk agar diffusion test results for fresh clinical isolates. For the stock isolates, AMS mean integer MICs approximated microdilution mean integer MICs with AMS, providing excellent MIC replicability. For ampicillin and cephalothin, 50 and 46.8%, respectively, of AMS integer MICs were within +/- 1 microgram/ml of the reference values, and 89.3 and 63.1% of AMS integer MICs were within +/- 2 micrograms/ml of the reference values. For the fresh clinical isolates, AMS and reference results were in disagreement for 4.5% of the antimicrobial agents tested, with 2.3% as a combination of "major" and "very major" errors.     

Direct identification and susceptibility testing by the AutoMicrobic system of gram-negative bacilli from urine specimens.

A total of 1,800 urine specimens were screened by Gram stain to detect bacteriuria. Pellets of bacteria were obtained by centrifuging specimens containing greater than or equal to 1 gram-negative bacillus of a single morphological type per oil immersion field. Direct susceptibility tests and identifications were performed from pellets by using the AutoMicrobic system (AMS). Results were compared with culture results by routine AMS methods. Of the 145 specimens showing only gram-negative bacilli on Gram stain, 113 grew greater than or equal to 10(5) CFU of a single species per ml. Compared with routine AMS identifications, the direct method correctly identified 105 (92.9%) of the isolates. Identifications were available within 8 h for 77% of the isolates. When compared by MICs, 93.2% of the direct susceptibility test results agreed with routine AMS results within one twofold dilution. Comparisons by category call indicated that overall complete and essential agreements were 89.9 and 97.8%, respectively, with 1.0% very major, 1.0% major, and 8.1% minor errors. Cefamandole and cephalothin had the lowest correlations by both comparisons. Within 8 h, susceptibility results were available for 94.3% of the isolates. This method offers the advantage of rapid detection, prompt processing, and earlier reporting of complete results for positive urine specimens.     

Rapid identification and antimicrobial susceptibility testing of gram-negative bacilli from blood cultures by the AutoMicrobic system.

A procedure was developed which allows direct identification and antimicrobial susceptibility testing of fermentative and nonfermentative gram-negative bacilli from positive blood cultures. A 10-ml sample was removed from turbid blood culture bottles, and the bacteria were washed and concentrated by centrifugation. The bacterial pellet was used to inoculate an Enterobacteriaceae Plus Identification Card and a Gram-Negative General Susceptibility Card of the AutoMicrobic system. Results with these cards were compared with results obtained with standard technique for 196 blood cultures seeded with recent clinical isolates. Identification of most cultures was available in 8 h, whereas the antimicrobial susceptibility results were available in an average of 4.7 h for all organisms. Direct identification was correct for 95% of the cultures, whereas the antimicrobial susceptibility data had an average agreement of 87% with 3.8% very major and 1.4% major errors. In using this procedure it was possible to provide accurate preliminary identification and results of antimicrobial susceptibility tests for gram-negative bacilli on the same day that a blood culture was determined to be positive.     

Evaluation of the Autoscan Walkaway system for rapid identification and susceptibility testing of gram-negative bacilli.

We evaluated the performance of the Autoscan Walkaway (W/A) system (MicroScan, Sacramento, Calif.) in conjunction with the fluorometric Neg Combo panels for rapid identification and susceptibility testing of gram-negative bacilli. Fermentative and nonfermentative gram-negative bacilli were tested in parallel with the W/A system and the Cathra Repliscan replicator (C/R) system (Cathra, St. Paul, Minn.). Conventional biochemical testing and agar dilution testing were used to resolve the identification and susceptibility testing discrepancies. Of 495 clinical isolates tested, 445 (90%) were correctly identified by the W/A system and 483 (98%) were correctly identified by the C/R system. Repeat testing by using updated versions of the W/A system's computer identification software failed to demonstrate improved identification accuracy. For susceptibility testing, the W/A system demonstrated 5.6% total interpretative category errors, including only 0.9% major and very major errors. The comparative C/R system produced only 1% errors overall, including 0.2% major and very major errors. Although the W/A system is highly automated and is capable of producing results rapidly, our findings suggest that additional identification and susceptibility testing refinements are needed before the system will be suitable for routine use.     

Evaluation of four anti-microbic susceptibility testing systems for gram-negative bacilli

A total of 200 clinical isolates were assayed by five anti-microbic susceptibility testing systems. Two frozen minimal inhibitory concentration (MIC) systems (MicroScan and Pasco), an automated MIC system (AMS, Vitek Systems), and the standard disk diffusion were compared with a reference broth dilution method. Organisms tested included 100 resistant clinical stock strains and 100 fresh random clinical isolates. Overall, there were 1,600 anti-microbic-organism combinations analyzed. The Pasco and MicroScan systems had no major discrepancies, the AMS system had seven, and the disk diffusion two. The number of very major discrepancies were as follows: AMS, 11; disk diffusion, 9; MicroScan, 5; Pasco, 2. Of the total 36 major or very major discrepancies in the study, 33% (12 of 36) were with an aminoglycoside and 44% (16 of 36) occurred with a second-generation cephalosporin, of which 10 of 16 were with cefamandole. Overall, there was a greater than 98.8% essential agreement with all systems compared with the reference method.     

Evaluation of the autoSCAN-W/A rapid system for identification and susceptibility testing of gram-negative fermentative bacilli.

The autoSCAN-Walk-Away (W/A) system for identification and susceptibility testing was evaluated for 400 gram-negative fermentative bacteria by using the API 20E (366 isolates) and/or tube biochemical tests as the reference identification system and a frozen microdilution MIC tray system for susceptibility testing. The W/A system performed well for identification of this group of organisms representing 14 genera and 30 species, showing a sensitivity of 96% and results available in 2 h. Of the 16 misidentifications, 6 were with Serratia liquefaciens. A total of 63 isolates (17%) required further tests to complete the identification, compared with 106 (29%) of the isolates which required additional tests for the API 20E identification. Approximately half (32) of the additional tests with the W/A system were required in order to separate Citrobacter diversus from C. amalonaticus. For susceptibility determinations, the W/A system demonstrated an overall agreement of 93% (4,102 determinations) with 40 major errors (0.98%). However, of the 906 resistant organism-drug combinations in the study, there were 115 very major errors, for a false-susceptibility rate of 12.7% of the resistance determinations. Among these very major errors, 80% occurred with piperacillin and the cephalosporins. The W/A system completed the MIC determinations in 7 h; however, the difficulty in detecting resistance with some antimicrobial agents limited the advantages of the rapid susceptibility testing.     

Evaluation of the AutoMicrobic system for identification of glucose-nonfermenting gram-negative rods.

With the Enterobacteriaceae Plus Biochemical Card, the AutoMicrobic system (Vitek Systems, Inc., Hazelwood, Mo.) is purported to identify members of the family Enterobacteriaceae and seven species of glucose-nonfermenting gram-negative rods: Pseudomonas aeruginosa, P. fluorescens, P. putida, P. maltophilia, P. cepacia, and Acinetobacter calcoaceticus (saccharolytic and non-saccharolytic). The latter capability was examined in this investigation. Of 410 glucose-nonfermenting rods included in the identification profile, the AutoMicrobic system correctly identified 366 (89.3%). Of 62 glucose-nonfermenting organisms not included in the identification profile, 41 (66%) were correctly reported as "unidentified organism." The usefulness of the AutoMicrobic system-generated identification probability in establishing criteria for acceptance of identification is discussed.     

Evaluation of autoSCAN-W/A and the Vitek GNI+ AutoMicrobic system for identification of non-glucose-fermenting gram-negative bacilli

The autoSCAN-W/A (W/A; Dade Behring Microscan Inc., West Sacramento, Calif.) and Vitek AutoMicrobic System (Vitek AMS; bioMérieux Vitek Systems, Inc., Hazelwood, Mo.) are both fully automated microbiology systems. We evaluated the accuracy of these two systems in identifying nonglucose-fermenting gram-negative bacilli. We used the W/A with conventional-panel Neg Combo type 12 and Vitek GNI+ identification systems. A total of 301 isolates from 25 different species were tested. Of these, 299 isolates were identified in the databases of both systems. The conventional biochemical methods were used for reference. The W/A correctly identified 215 isolates (71. 4%) to the species level at initial testing with a high probability of >/=85%. The Vitek GNI+ correctly identified 216 isolates (71.8%) to the species level at initial testing with a high probability of >/=90%. After additional testing that was recommended by the manufacturer's protocol, the correct identifications of the W/A and Vitek GNI+ improved to 96.0 and 92.3%, respectively. The major misidentified species were Sphingomonas paucimobilis and Agrobacterium radiobacter in the W/A system and Acinetobacter lwoffii, Chryseobacterium indologenes, and Comamonas acidovorans in the Vitek GNI+ system. The error rates were 4.0 and 7.6%, respectively. The overall accuracy for both systems was above 90% if the supplemental tests were applied. There was no significant difference in accuracy (P > 0.05) between the two systems.     

Rapid automated identification of gram-negative bacilli from blood cultures with the AutoMicrobic system.

Automated identification of gram-negative bacilli directly from blood culture bottles by using the AutoMicrobic System (AMS) was evaluated with a modified procedure for the AMS Enterobacteriaceae-plus nonfermenter identification card. A total of 150 strains were tested (44 clinical and 106 seeded) and compared with a conventional identification procedure. These strains included 107 Enterobacteriaceae and 43 oxidase-positive or glucose-nonfermenting, or both, organisms. AMS identifications on one of these strains were not interpretable owing to equal probability AMS identification values. Of the remaining 149 strains, 138 (92.6%) were correctly identified within 8 to 13 h of the first reading. Of 69 identifications analyzed after 6 h of incubation, 91% were correct. This procedure was found to be rapid, convenient, and nonlabor intensive and is recommended for presumptive identification of gram-negative bacilli in blood cultures.     

Comparison of the Cathra Repliscan II, the AutoMicrobic system Gram-Negative General Susceptibility-Plus Card, and the Micro-Media System Fox Panel for dilution susceptibility testing of gram-negative bacilli.

A comparative evaluation was done to test the accuracy of the Cathra Repliscan II agar dilution system (Diagnostic Equipment, Inc., St. Paul, Minn.), the AutoMicrobic system with Gram-Negative General Susceptibility-Plus Card (Vitek Systems, Inc., Hazelwood, Mo.), and the Micro-Media Fox Panel micro broth dilution system (Micro-Media Systems, Inc., San Jose, Calif.) in determining MICs of 12 antibiotics for 200 gram-negative bacilli. Of the 200 strains tested, 12 isolates did not grow in one of the three systems. The 188 remaining organisms included 158 members of the family Enterobacteriaceae, 20 Pseudomonas spp., 5 Acinetobacter sp., 3 Aeromonas spp., and 2 Vibrio spp. A total of 2,256 organism-antibiotic combinations were analyzed for each system. An MIC was considered correct if two of the three systems were in agreement. When disagreements occurred, correct MICs were determined by the standard agar dilution method. With this criterion, overall agreements of the Cathra Repliscan II system, AutoMicrobic system, and Micro-Media Fox Panel system were 94.7, 94.9, and 95.5%, respectively. Tetracycline (20%), nitrofurantoin (20%), and ampicillin (16%) accounted for 56% of the discrepancies observed. These results indicate that all three systems perform with a high degree of accuracy for susceptibility testing of gram-negative bacilli.     

Comparison of the AutoMicrobic system and a conventional tube system for identification of nonfermentative and oxidase-positive gram-negative bacilli

The AutoMicrobic system (AMS) Enterobacteriaceae-plus Biochemical Card was developed to identify a select group of 10 species of glucose-nonfermentative and oxidase-positive fermentative gram-negative bacilli. In this study, 159 nonenteric clinical isolates were identified by the AMS and conventional tube biochemicals based on E. O. King's (Centers for Disease Control) identification schema. The AMS properly identified 96.7% (117 of 121) of isolates whose taxa were included in the AMS data base. Of 38 isolates (94.7%) in which taxa were not included in the data base, 36 were correctly called unidentified organisms. A principal advantage of the AMS is the automated identification of frequently isolated nonenterics in a period of only 8 to 13 h. The AMS, with the use of the Enterobacteriaceae-plus Biochemical Card appears to be a rapid and accurate system for the identification of the most commonly isolated nonfermentative and oxidase-positive fermentative gram-negative bacilli.     

Evaluation of the VITEK 2 system for rapid direct identification and susceptibility testing of gram-negative bacilli from positive blood cultures

This study explores the possibility of combining the BacT/Alert Microbial Detection System with the VITEK 2 system to achieve rapid bacterial identification and susceptibility testing. Direct inoculation of bacterial suspension to the VITEK 2 ID-GNB card and AST-NO09 card was made by differential centrifugation of blood cultures of organisms with gram-negative enteric bacillus-like morphology. A total of 118 strains were investigated; of these, 97 (82.2%) strains were correctly identified to the species level and 21 (17.8%) strains were not identified; by comparing the results with those of the reference method of API identification systems using a pure culture, it was found that no strain had been misidentified. Among the 21 strains with no identification, 13 (61.9%) strains were nonfermenters. The direct-identification reporting time of VITEK 2 was 3.3 h. Direct testing of susceptibility to 11 antibiotics, i.e., amikacin, cefepime, ceftazidime, ciprofloxacin, gentamicin, imipenem, meropenem, netilmicin, piperacillin, piperacillin-tazobactam, and tobramycin, was also performed by using the broth microdilution (MB) method according to the NCCLS guidelines as a reference. After comparing the MICs of the VITEK 2 system with those obtained by the MB method within +/-twofold dilution, it was determined that the 1,067 organism-antibiotic combinations had an overall correct rate of 97.6% (1,041 combinations). The rates of susceptibility to the 11 antibiotics ranged from 88.7 to 100%, respectively. Only two (0.2%) and four (0.4%) combinations of the susceptibility tests gave very major errors (i.e., reported as sensitive by the VITEK 2 system but shown to be resistant by the MB method) and major errors (i.e., reported as resistant by the VITEK 2 system but shown to be sensitive by the MB method), respectively. The reporting time for the direct testing of susceptibility against the 11 antibiotics for 97 blood culture isolates by the VITEK 2 system ranged from 3.3 to 17.5 h. Compared with conventional methods that require 1 or 2 days, this method can make same-day reporting possible and thus permit better patient management.     

Comparison of three automated systems for antimicrobial susceptibility testing of gram-negative bacilli.

Several instruments for automated or semiautomated susceptibility testing are currently available. Three of these instruments, Autobac (General Diagnostics, Warner-Lambert Co., Morris Plains, N.J.), MS-2 (Abbott Laboratories, Dallas, Tex.), and AutoMicrobic system (AMS) (Vitek, Inc., Hazelwood, Mo.) were compared for antimicrobial susceptibility testing of gram-negative bacilli. A total of 207 isolates representing 29 species of gram-negative bacilli were tested simultaneously by each instrument and by a standardized disk diffusion reference method. Nine antimicrobial agents, including ampicillin, carbenicillin, cephalothin, gentamicin, tobramycin, amikacin, tetracycline, trimethoprim-sulfamethoxazole, and nitrofurantoin were tested. Discrepancies between the results of the automated and reference disk diffusion methods were resolved by agar dilution testing. Overall, 93% of the Autobac and MS-2 results and 83% of the AMS results were in agreement with the results obtained by the reference methods. The results of the Autobac, MS-2, and AMS systems respectively included 3.3, 2.3, and 4.2% major and very major discrepancies. Excessive testing discrepancies were found for certain drugs, including ampicillin, tetracycline, and nitrofurantoin, and for certain organisms, including species of Providencia, Serratia, and Citrobacter. The results of this comparison of three automated systems for antimicrobial susceptibility testing indicate that the Autobac and MS-2 instruments provided highly reliable results. The AMS need further development of its susceptibility testing capability to eliminate an unacceptably high number of minor discrepancies.     

Evaluation of the Merlin MICRONAUT system for rapid direct susceptibility testing of gram-positive cocci and gram-negative bacilli from positive blood cultures

Bloodstream infections are life-threatening conditions which require the timely initiation of appropriate antimicrobial therapy. We evaluated the automated Merlin MICRONAUT system for rapid direct microtiter broth antimicrobial susceptibility testing (AST) of gram-positive cocci and gram-negative bacilli from BACTEC 9240 bottles with positive blood cultures in comparison to the standard method for the Merlin MICRONAUT system. This prospective study was conducted under routine working conditions during a 9-month period. Altogether, 504 isolates from 409 patients and 11,819 organism-antibiotic combinations were evaluated for comparison of direct and standard AST methods. For gram-negative bacilli, direct and standard AST of 110 isolates was evaluated and MIC agreement was found for 98.1% of 2,637 organism-antibiotic combinations. Category (susceptible, intermediate susceptible, resistant [SIR]) agreement was found for 99.0%, with results for 0.04% of combinations showing very major errors, those for 0.2% showing major errors, and those for 0.8% showing minor errors. For gram-positive cocci, 373 isolates were evaluated and MIC agreement was found for 95.6% of 8,951 organism-antibiotic combinations. SIR agreement was found for 98.8%, with results for 0.3% of combinations showing very major errors, those for 0.4% showing major errors, and those for 0.5% showing minor errors. Although the number of tested isolates was limited (n = 33), direct AST of streptococci was performed for the first time, yielding promising results with SIR agreement for 98.6% of 363 organism-antibiotic combinations. In conclusion, direct AST of gram-negative bacilli and gram-positive cocci from positive blood cultures with the MICRONAUT system is a reliable technique that allows for the omission of repeat testing of subcultured isolates. Thereby, it reduces the time to results of blood culture testing and may have a positive impact on patient care.     

Comparison of the autoSCAN-W/A rapid bacterial identification system and the Vitek AutoMicrobic system for identification of gram-negative bacilli

The autoSCAN-W/A (W/A; Baxter MicroScan, West Sacramento, Calif.) with the new fluorometric Rapid Neg Combo 1 (RNC) panel is a fully automated fluorometric system for identification of both enteric and nonenteric gram-negative bacilli within 2 h. We compared the W/A with the Vitek AutoMicrobic System (Vitek AMS; Vitek Systems, Inc., Hazelwood, Mo.) for identification of 383 clinical isolates of gram-negative bacilli. The API 20E (Analytab Products, Plainview, N.Y.) and conventional biochemical testing were used as the reference systems. The W/A correctly identified 336 isolates (87.7%) to the species level and classified an additional 29 isolates (7.6%) as correct with low probability (overall identification = 95.3%); the Vitek AMS correctly identified 355 isolates (92.7%) to the species level and classified an additional 8 isolates (2.1%) as correct with low probability (overall identification = 94.8%). A common set of 134 isolates of gram-negative bacilli was tested in both participating laboratories as a means of assessing interlaboratory agreement with both the W/A and the Vitek AMS. The overall agreements between the two laboratories were 86% with the W/A and 92% with the Vitek AMS. The W/A performed comparably to the Vitek AMS for identification of most gram-negative bacilli, actually exceeding the Vitek AMS for identification of nonenteric bacilli. Rapid time to identification and a high level of automation make the W/A an attractive system for clinical microbiology laboratories.     

Evaluation of PASCO MIC-ID system for identifying gram-negative bacilli.

A production model of the semi-automated PASCO MIC-ID system (PASCO Laboratories, Wheat Ridge, Colo.) was evaluated with 122 groups or species of gram-negative bacilli, which included typical (499 cultures) and atypical (37 cultures) strains of fermenters and nonfermenters. The PASCO system identified 90.9% of 536 cultures accurately; these included 90.8% of 152 nonfermenters, 93.8% of 308 enteric fermenters, and 78.9% of 76 oxidase-positive fermenters. These results were obtained with the aid of serologic tests and a few additional biochemical tests, when recommended by the PASCO system. Of the 14 misidentified nonfermenters, 3 were Pseudomonas paucimobilis, 3 were Weeksella virosa (Centers for Disease Control group IIf), 2 were Xanthomonas (Pseudomonas) maltophilia, and 6 were randomly distributed among the other groups and species tested. The 19 enteric fermenters that were misidentified were randomly distributed among the groups and species tested. Of the 16 misidentified oxidase-positive fermenters, 4 were Pasteurella ureae, and 12 were randomly distributed among the other groups and species. The system identified the most commonly encountered organisms at a rate of 95% or better. The PASCO system is easy to inoculate and read. A slightly improved data base should remedy most of the identification problems.     

Comparison of two automated instrument systems for rapid susceptibility testing of gram-negative bacilli.

The Vitek AutoMicrobic System with GSC-plus cards and the Abbott MS-2 system were tested in parallel and the results were compared directly with those of a reference microdilution minimal inhibitory concentration (MIC) procedure on a group of 262 clinical isolates of the family Enterobacteriaceae and of Pseudomonas aeruginosa. Results of both systems were compared with the reference MIC for category agreement, and in addition, the Vitek MICs were compared with those obtained by the reference procedure. The Vitek system provided an essential category correlation of 89.4% for enteric bacteria and 97.0% for P. aeruginosa. Vitek MICs agreed within 1 twofold dilutional increment for 86.3% of the enteric bacteria tested and for 96.2% of the P. aeruginosa isolates. The Abbott MS-2 essential categoric agreement was 92.0% for enteric bacteria and 92.4% for P. aeruginosa. If only aminoglycosides or carbenicillin were considered for P. aeruginosa isolates, the essential category agreement was 92.5% for the Vitek and 93.3% for the MS-2. The majority of MS-2 category errors (13 of 19) with P. aeruginosa involved gentamicin results on isolates whose reference MICs were 8 micrograms/ml and whose MS-2 results were susceptible (MIC less than or equal to 4 micrograms/ml). Retesting of the P. aeruginosa isolates in calcium-supplemented MS-2 broth increased the essential agreement for the aminoglycosides to 97.5%.     

Antimicrobial resistance among gram-negative bacilli to newer aminoglycosides and beta-lactams.

Two hundred and fifty nine i.e. 30.72% of 843 strains of gram negative bacilli were resistant to gentamicin. When tested against newer aminoglycosides, 48.9% were resistant to tobramycin, 4.1% to amikacin and 20.5% to netilmicin. Marked cross-resistance between gentamicin and tobramycin was seen. All these strains were resistant to ampicillin, amoxycillin and co- trimoxazole. However, a high level of antimicrobial activity was seen with third generation cephalosporins, cefotaxime and ceftriaxone. Minimum inhibitory concentrations of gentamicin and tobramycin were of significant level. Gentamicin resistance could be transferred directly in 20 of 45 strains of S. typhimurium where conjugation experiments were done.     

Critical evaluation of the AutoMicrobic system gram-negative identification card for identification of glucose-nonfermenting gram-negative rods.

During a 6-month study we critically evaluated the accuracy of the AutoMicrobic system Gram-Negative Identification Card (Vitek Systems, Inc., Hazelwood, Mo.) in identifying glucose-nonfermenting gram-negative bacilli by testing 419 selected isolates in parallel with a conventional reference method. Of 356 isolates included in the AutoMicrobic system profile, a total of 307 (86.2%) were correctly identified, 36 (10.1%) were not identified, and 13 (3.7%) were misidentified. Fifty-eight of 63 (92%) isolates not included in the profile were correctly reported as "unidentified organisms." Overall, if the first-choice identification was always accepted, only 18 (4.3%) isolates would have been incorrectly reported. When first-choice identifications appended with the special message "questionable biopattern" were rejected, and organisms were screened for characteristic odor and antimicrobial susceptibility before final acceptance of the AutoMicrobic system report, the number of misidentifications was reduced to 5 (1.2%). The average time to identification with the AutoMicrobic system Gram-Negative Identification Card was 15 h. This compares favorably with the 65 h required by the reference method.     

Comparison of Crystal Enteric/Nonfermenter system, API 20E system, and Vitek AutoMicrobic system for identification of gram-negative bacilli.

A comparative evaluation of the Crystal Enteric/Nonfermenter system (Crystal; Becton Dickinson, Cockeysville, Md.), API 20E (API; bioMérieux Vitek, Inc., Hazelwood, Mo.), and the Vitek GNI card (Vitek; bioMérieux Vitek) was performed with 512 clinical isolates of gram-negative bacilli, including 381 members of the family Enterobacteriaceae and 131 nonenteric bacilli. With supplemental testing, API, Crystal, and Vitek correctly identified to the genus and species level 505 (98.6%), 489 (95.5%), and 494 (96.5%) of the 512 isolates, respectively. Supplemental testing, as specified by the manufacturer, was required to identify 119 (23.2%), 18 (3.5%), and 5 (1.0%) of the isolates with the three systems, respectively. Of the 381 isolates from the family Enterobacteriaceae, API and Crystal correctly identified 90.3 and 91.6% by 18 to 24 h without supplemental testing, respectively, and Vitek identified 92.4 and 96.1% following 10 and 18 h of incubation, respectively. Of the 131 nonenteric organisms, API and Crystal correctly identified 28.2 and 93.9% by 18 to 24 h without supplemental testing, respectively, and Vitek identified 84.0% by 10 h and 93.9% by 18 h. Errors in identification with each system were infrequent and appeared to be randomly distributed among the genera evaluated. The three systems were comparable in accuracy when either a weighted clinical laboratory profile of organisms or a group of selected isolates in a stress test sample was evaluated (P > 0.05). There were no significant differences between the three systems in their ability to identify either the isolates in the weighted group or those in the stress test (P > 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)