Annual Meeting of the American Society for Microbiology Paper C-250 May, 2003, Washington, DC Analysis of the Comparative Workflow of the VITEK 2 and the System U. Eigner, A. Caganic, U. Wild, D. Bertsch, A.-M. Fahr Laboratory Group Heidelberg, Germany. Revised Abstract Background The implementation of automated bacterial identification (ID) and antimicrobial susceptibility (AST) test systems in routine microbiology lab. provides the clinical benefit of accurate results in an efficient time frame. According to this topic we compared the workflow of the automated (biomérieux, Marcy L Etoile, France) with the system (BDDiagnostic System, Md., Sparks, USA) measuring every manipulation step setting up the different devices and the analysis time for AST. Material and Methods A total of 307 clinical isolates was tested, with 144 Gram-positive and 163 Gram-negative strains. The isolates were randomly collected in our routine lab. from various specimens. Time measurement of the different preparation steps was performed and the analysis time of AST results collected. Discrepant ID results were compared to the API systems (biomérieux). The results of AST were compared to the reference broth microdilution method (Biotest AG, Dreieich, Germany) Results The mean manipulation time for one isolate was 1mn 31s for the and 2mn 59s for the instrument. Concerning AST the average time to result for the and the system for Gram-positive cocci was 8h 53mn and 12h 14mn, respectively. Results for Gram-negative rods were available in an average of 8h 02mn with and in 12h 01min with the system. 98%/ 98% of the Gram-positive cocci and 98%/ 97% of the Gram-negative rods were correctly identified by the and the system, respectively. After applying the expert rules category agreement (CA) was 99,3% and 98,9% for Gram-positive cocci and 95,0% and 95,2% for Gram-negative rods for the and the system. Conclusions Both systems showed excellent results for ID and AST performance in comparison to the reference method. manipulation time for one isolate was reduced (1mn 31s) in comparison to (2min 59) due to a higher degree of automation. Introduction and Purpose Only a few workflow studies had been performed comparing the instrument with manual methods (1,2,4) and two studies compared the with the automated instrument Walkaway (Dade-MicroScan, W. Sacramento, Calif.) (3,5). The main objective of this study was to compare the workflow of the system with the BD Automated Microbiology System. Additionally the time-to-result and the accuracy of identification (ID) and antimicrobial susceptibility test results (AST) were evaluated. Materials and Methods Bacterial isolates. A total of 307 clinical strains from various specimens were tested, including the following isolates: Enterobacteriaceae (141), Nonfermenters (22), staphylococci (93), and enterococci (51) (Table 1). For all 307 isolates time-to-result and ID/AST performance was tested. Manipulation time was measured in batches, each with 7 isolates, for a total of 39 batches, comprising 273 isolates. Workflow. Two skilled technicians were involved in this study. Manipulation time: The steps are included in Table 3. Using a stopwatch, all steps were timed by an observer not involved in testing process. Maintenance of the instruments: The time to perform the regular maintenance (daily, weekly, monthly) required for each system was recorded. Time-to-result: The time to result was calculated using the inoculation time indicated on the laboratory reports. Antimicrobials. The antimicrobials tested and the cards and panels used in this study are shown in Table 2. Reference system. For discrepant ID results between the and devices, an ID API 20 E, API 32 Staph, and API 32 Strep (biomérieux) were set up in duplicate. For discrepant AST results, frozen microdilution Page 1
panels were set up for each isolate according to NCCLS guidelines (6). Very major errors (VME), major errors (ME), and minor errors (me) > 10% were repeated in duplicate in all systems. A majority rule determined the resolved outcome. Data analysis. Workflow: The average time to perform each step was calculated over the 39 batches. Statistical analysis was performed using the Student s t-test statistic to determine if a significant difference existed between manipulation times for both systems. AST results: The final results were calculated after applying the expert rules of the respective instrument. The following parameters were calculated: Category agreement (CA), the concordance of SIR results of the test systems to the reference system; VME, false Table 1. Isolates Tested susceptible or results; ME, false resistant or results; me, S or R in test system, I in reference system, and vice versa. The error rates were calculated as follows: No _ me % me = Total _ isolates _ tested No _ ME % ME = Total _ S _ isolates No _ VME % VME = Total _ R _ isolates AST Time to Result: Statistical analysis was performed using the Student s t-test statistic to determine if a significant difference existed between times to result for both systems. ID: Percentages of concordant and discordant results were compared after resolution with a comparator system. Table 2. Antibiotics and Test Panels Bacterial isolates n Bacterial isolates n Enterobacteriaceae 141 Staphylococci 93 Escherichia coli 30 Staphylococcus aureus 49 ESBL- 15 MSSA 31 ESBL+ 11 MRSA 18 AmpC+ 4 S. epidermidis 22 Klebsiella spp. 36 S. haemolyticus 8 ESBL- 26 S. lugdunensis 2 ESBL+ 11 S. simulans 3 Proteus spp. 16 S. sciuri 1 Providenica spp. 5 S. saprophyticus 1 Serratia spp. 15 S. capitis 1 Citrobacter spp. 9 other CoNS 6 Enterobacter spp. 17 MSCoNS 27 Morganella morganii 8 MRCoNS 18 other Enterobact. 5 Nonfermenters 22 Enterococci 51 Pseudomonas spp. 14 E. faecalis 50 Acinetobacter spp. 8 E. faecium 1 Enterobacteriaceae/Nonfermenters Staphylococci Enterococci : AST-N021/ ID-GNB : NMIC/ID 14 : AST-P 523/ P524/ ID-GPC : NMIC/ID 15 Ampicillin Penicillin Vancomycin Piperacillin Oxacillin Teicoplanin Piperacillin/Tazobactam Erythromycin Tetracycline Meropenem Clindamycin Cefazolin Vancomycin Cefuroxime Teicoplanin Cefotaxime Rifampicin Ceftazidime Ofloxacin Cefepime Gentamicin Levofloxacin Tetracycline Cotrimoxazol Results Workflow. The results of the different steps of Time-to-result data are included in Table 5. the manipulation time are shown in Tables 3 For each bacterial group and overall, the mean and 3A. The steps associated with the time-to-result was significantly greater for workflow of each system were timed for thirtynine when compared to. Analysis batches of 7 organisms. A statistical of the time-to-result differences from pairing by analysis of the batch time reveals that the organism for all strains indicated that the mean time of 1251 seconds per batch for overall mean difference between results was 3 was significantly longer than the 635 hours and 40 minutes longer for. A seconds per batch for. This t-test significant time savings of 4 hours and 10 result, with a p-value <0.001, provides minutes for Enterobacteriaceae, 2 hours and 52 evidence of an approximate savings of 616 minutes for nonfermentors, 4 hours and 18 seconds, or over 10 minutes, per batch. While minutes for Staphylococci, and 1 hour and 34 the direct conversion does not appropriately for Enterococci was discovered with. take into account all of the contributing factors, ID. The results are shown in Table 6. the difference in this data suggests an AST. The AST-results of all antimicrobials estimated savings of 1 minute and 28 seconds tested and evaluated in this study are shown in per organism. The results of the maintenance Table 7. of the instruments are included in Table 4. Page 2
Table 3. Average manipulation time for steps with and systems (in sec per batch = 7 isolates) Step Time/ sec PHOENIX Time/ sec 1 Organize bench 19 Organize bench 19 2 Label purity plates 15 Label purity plates 16 3 Label tubes with saline* 44 Label ID-broth* 56 4 Dispense saline 12 Drop indicator 67 Make suspension 5 230 Check with Densichek/vortex Inoculate purity plate 47 Open package 63 Inoculate ID Broth Check with Crystalspec/vortex 462 Enter accession No./ Read barcode Put card in rack 138 Transfer 25 µl to AST broth 137 6 Open package Label panels Place ID and AST Broth in inoculation station Inoculate panels with ID and AST broth 107 96 Close the panels 32 Put panels in transport caddy 17 Inoculate purity plates 67 10 Put carrier in V2 9 Enter accession no. Put panels in PHX 113 11 Put purity plates in incubator 46 Put purity plates in incubator 46 12 Remove panels from instrument and throw away 13 Remove panels from instrument and throw away 17 Average per batch 635 1251 Average per isolate 1 mn 31 s 2 mn 59 s Table 3A. Overall Workflow Time Comparison (in seconds, with 7 organisms per batch) System Number of Batches Mean Minimum Maximum P-value PHOENIX 39 1251 1052 1494 39 635 539 783 < 0.001 Page 3
Table 4. Instrument Maintenance Interval daily Step Time (s) Step Time (s) temperature 15 Calibration of CrystalSpec 25 every three days Fiiling of tips/ Changing of NaCl bags (for V2 120) 120 weekly Cleaning of optic system 420 Calibration of DensiCheck 11 monthly signal lamps, acoustic-, visual alarm 73 Cleaning carousel, boats, optical devices 900 Table 5. AST Time to Result Comparison (h:mm) Bacterial Statistical Value PHOENIX Paired Difference N 140 141 140 Enterobacteriaceae Mean 7:31 * 11:41 * 4:10 ** Min. - Max. 5:30-13:00 7:50 15:59 N 22 22 22 Nonfermenter Mean 11:19 * 14:11 * 2:52 ** Min. - Max. 7:00 17:45 8:40 16:00 N 89 93 89 Staphylococci Mean 8:34 * 12:55 * 4:18 ** Min. - Max. 6:15 18:00 7:11 16:01 N 51 51 51 Enterococci Mean 9:26 * 11:00 * 1:34 ** Min. - Max. 6:00 12:15 7:01 15:58 N 302 307 302 Combined Mean 8:26 * 12:07 * 3:40 ** Min. - Max. 5:30 18:00 7:01 16:01 * Significantly different mean values, > ; all t-test p-values < 0.001 ** Mean difference significantly > 0 hours; all t-test p-values < 0.001 Table 6. Identification Results for and n Bacterial groups isolates tested concordant % discordant % concordant % discordant % Enterobacteriaceae 140 a 98 c 2 96 d 4 Nonfermenters 22 100 0 100 0 Staphylococci 92 a 97 3 97 3 Enterococci 51 100 0 100 b 0 a one isolate not solved by reference method b motility test solved 2 discrepancies c 3 isolates incorrect for genus d 3 isolates incorrect for genus, 2 isolates incorrect only for species Page 4
Table 7. AST Results for and Bacterial Groups Enterobacteriaceae Nonfermenters Staphylococci Enterococci Antibiotics S I R total me ME VME %me %ME %VME %CA me ME VME %me %ME %VME %CA Ampicillin 24 4 113 141 1 0.7 0.0 0.0 99.3 3 1 2.1 4.2 0.0 97.2 Piperacillin 90 11 40 141 4 2 2.8 0.0 5.0 95.7 4 2.8 0.0 0.0 97.2 Piperacillin/Tazobactam 121 8 12 141 11 7.8 0.0 0.0 92.2 7 1 5.0 0.8 0.0 94.3 Cefazolin 42 7 92 141 9 4 6.4 0.0 4.3 90.8 4 2 2.8 0.0 2.2 95.7 Cefepime 134 1 6 141 2 1 1.4 0.0 16.7 97.9 3 2.1 0.0 0.0 97.9 Cefuroxim 60 16 65 141 3 2.1 0.0 0.0 97.9 0.0 0.0 0.0 100.0 Cefotaxime 124 4 13 141 5 3.5 0.0 0.0 96.5 8 5.7 0.0 0.0 94.3 Ceftazidime 125 4 12 141 6 4.3 0.0 0.0 95.7 5 3.5 0.0 0.0 96.5 Meropenem 141 0 0 141 0.0 0.0 0.0 100.0 0.0 0.0 0.0 100.0 Ciprofloxacin 113 3 25 141 3 2.1 0.0 0.0 97.9 7 5.0 0.0 0.0 95.0 Levofloxacin 114 9 18 141 13 9.2 0.0 0.0 90.8 9 6.4 0.0 0.0 93.6 Gentamicin 120 4 17 141 5 3.5 0.0 0.0 96.5 6 4.3 0.0 0.0 95.7 total 1208 71 413 1692 62 0 7 3.7 0.0 1.7 95.9 56 2 2 3.3 0.2 0.5 96.5 Piperacillin 13 2 7 22 1 0.0 0.0 14.3 95.5 1 4.5 0.0 0.0 95.5 Piperacillin/Tazobactam 18 0 4 22 1 4.5 0.0 0.0 95.5 2 9.1 0.0 0.0 90.9 Cefepime 17 3 2 22 4 18.2 0.0 0.0 81.8 5 22.7 0.0 0.0 77.3 Cefotaxime 5 6 11 22 3 13.6 0.0 0.0 86.4 4 1 18.2 0.0 9.1 77.3 Ceftazidime 14 4 4 22 5 1 22.7 0.0 25.0 72.7 3 1 13.6 0.0 25.0 81.8 Meropenem 17 2 3 22 4 1 18.2 0.0 33.3 77.3 6 27.3 0.0 0.0 72.7 Ciprofloxacin 11 2 9 22 2 9.1 0.0 0.0 90.9 1 4.5 0.0 0.0 95.5 Levofloxacin 11 4 7 22 3 13.6 0.0 0.0 86.4 4 18.2 0.0 0.0 81.8 Gentamicin 16 0 6 22 1 4.5 0.0 0.0 95.5 2 9.1 0.0 0.0 90.9 total 122 23 53 198 23 0 3 11.6 0.0 5.7 86.9 28 0 2 14.1 0.0 3.8 84.8 Penicillin 31 0 62 93 0.0 0.0 0.0 100.0 0.0 0.0 0.0 100.0 Oxacillin 57 0 36 93 0.0 0.0 0.0 100.0 1 0.0 1.8 0.0 98.9 Vancomycin 93 0 0 93 0.0 0.0 0.0 100.0 0.0 0.0 0.0 100.0 Teicoplanin 93 0 0 93 0.0 0.0 0.0 100.0 1 1.1 0.0 0.0 98.9 Ofloxacin 55 0 38 93 0.0 0.0 0.0 100.0 1 1.1 0.0 0.0 98.9 Ciprofloxacin 56 0 37 93 1 0.0 1.8 0.0 98.9 0.0 0.0 0.0 100.0 Tetracycline 82 0 11 93 1 0.0 0.0 9.1 98.9 1 1 1.1 0.0 9.1 97.8 Gentamicin 74 0 19 93 1 1.1 0.0 0.0 98.9 1 1 1.1 1.4 0.0 97.8 Erythromycin 42 1 50 93 3 1 3.2 0.0 0.0 95.7 2 2.2 0.0 0.0 97.8 Clindamycin 63 0 30 93 0.0 0.0 0.0 100.0 1 1 1.1 1.6 0.0 97.8 Cotrimoxazol 81 11 1 93 1 1 0.0 1.2 0.0 97.8 1 0.0 1.2 0.0 98.9 Rifampicin 92 1 0 93 0.0 0.0 0.0 100.0 0.0 0.0 0.0 100.0 total 819 13 284 1116 4 2 3 0.4 0.2 1.1 99.2 7 4 1 0.6 0.5 0.4 98.9 Vancomycin 51 51 0.0 0.0 0.0 100.0 2 3.9 0.0 0.0 96.1 Teicoplanin 51 51 0.0 0.0 0.0 100.0 0.0 0.0 0.0 100.0 Tetracycline 19 32 51 0.0 0.0 0.0 100.0 1 2.0 0.0 0.0 98.0 total 217 0 38 255 0 0 0 0.0 0.0 0.0 100.0 3 1.2 0.0 0.0 98.8 Conclusions Manipulation time of 1min 31sec for one isolate was required with the VITEK 2 instrument compared to the automated system, which required 2min 59sec for one isolate. (p < 0.001) Overall, the AST results were obtained 3h 40mn faster (p < 0.001) with the instrument (Enterobacteriaceae 7h 31mn, Nonfermenters 11h 19mn, staphylococci 8h 34mn, enterococci 9h 26mn) compared to the instrument (Enterobacteriaceae 11h 41mn, Nonfermenters 14h 11mn, staphylococci 12h 55mn, enterococci 11h 00mn). ID and AST results were excellent for both instruments. References 1. Auckenthaler, R. 1997. Integrating automation in a microbiology laboratory. 7 th Eur. Cong. Clin. Microbiol. Infect. Dis. 1997, abstr. S 27. 2. Eigner, U., A. Caganic, D. Schumacher, R. Englert, A.M. Fahr. 1999. Workflow evaluation and performance of the new automated in comparison with BBL rapid crystal and agar diffusion. 99 th Gen. Meet. Am. Soc. Microbiol. 1999, abstr. C 85. 3. Goossens, W.H.F., H.J.A. Van Vliet, H.A. Verbrugh. 1999. Workflow analysis and evaluation of the accuracy of versus Walk a Way. 9 th Eur. Cong. Clin. Microbiol. Infect. Dis., abstr. P 0922. 4. Hosein, I., M. Heginbothom, A. Paull. 2001. Work-flow analysis of the automated identification and antimicrobial susceptibility testing system in a British clinical laboratory. 11 th Eur. Cong. Clin. Microbiol. Infect. Dis, abstr. P 1516. 5. Larone, D. H., Tucci, L. J., D. O. Samide. 2000. Time Study of Three Automated Systems for the Identification and Susceptibility of Bacteria: The MicroScan WalkAway 96, VITEK, and. 100 th Gen. Meet. Am. Soc. Microbiol., abstr. C-279. 6. National Committee for Clinical Laboratory Standards. 2002. Performance standards for antimicrobial susceptibility testing: Twelfth informational supplement, vol. 22, No. 1. NCCLS document M100- S12. Approved Standard M2-A5. National Committee for Clinical Laboratory Standards, Wayne, Pa. Page 5