J Vet Diagn Invest :164 168 (1998) Evaluation of a computerized antimicrobial susceptibility system with bacteria isolated from animals Susannah K. Hubert, Phouc Dinh Nguyen, Robert D. Walker Abstract. The BIOMIC is a computerized system used to calculate the minimal inhibitory concentration (MIC) of an antimicrobial from a zone of inhibition generated by a disk diffusion test. This system was developed using bacterial pathogens of human origin. This study investigated the use of the BIOMIC system for determining MICs for bacterial pathogens from animals. The MICs generated by the BIOMIC system were compared with the MICs generated using a broth microdilution testing method. A total of 663 drug organism combinations was tested. These combinations included 3 species of gram-positive bacteria, species of gramnegative bacteria, and the antimicrobial s ampicillin,,, ciprofloxacin, enrofloxacin, trimethoprim/sulfamethoxazole, tetracycline, and erythromycin. Overall, the MICs generated by the BIOMIC system correlated with the broth microdilution MICs for 72 of the total drug organism combinations tested. The Pseudomonas aeruginosa strains tested showed the highest between the 2 systems, with for all antibacterial s tested, whereas Pasteurella haemolytica, Pasteurella multocida, and enterococci showed the least (76, 7, and 47, respectively). Among these organisms, trimethoprim-sulfa showed the least (31) and ciprofloxacin showed the greatest (91). These results indicate that the BIOMIC system could be a useful tool in veterinary medicine for producing quantitative antimicrobial susceptibility results. However, it is currently unreliable for some drug bacteria combinations. This discrepancy possibly could be corrected by modification of the software using data points generated by a large-scale study. The importance of antibacterial chemotherapeutic s in the treatment of infectious disease processes is well established. Historically, a practitioner could select an effective drug based on clinical experience. However, with an increase in bacterial resistance to commonly used antibacterial s, it has become increasingly difficult for clinicians to empirically select an appropriate antibacterial. An alternative to empirically selecting an antibacterial is in vitro antimicrobial susceptibility testing on bacteria isolated from properly collected samples. There are basically 2 types of results that are provided by in vitro susceptibility testing: qualitative results, i.e., susceptible, intermediate, or resistant, and quantitative results, i.e., the minimum concentration of an antimicrobial required to inhibit the growth of the bacteria (MIC). Two common in vitro methods for determining susceptibility profiles of bacterial pathogens are disk diffusion testing and dilution testing. Disk diffusion testing involves applying a single concentration of an antimicrobial to a seeded agar medium and allowing the drug to diffuse into the surrounding medium. The bacteria on the agar medium are thus exposed to a continuous gradient of drug with the concentration diminishing as the distance from the disk increases. A From the Department of Microbiology, College of Veterinary Medicine, A3 Veterinary Medical Center, Michigan State University, East Lansing, MI 48824. Received for publication July 23, 1997. zone of inhibition forms when the concentration of drug reaches a concentration of bacteria it can no longer inhibit. The edge of this zone thus represents an MIC. However, because the concentration of the antibacterial at the edge of the zone of inhibition may vary for each drug bacterium combination, the MIC cannot be easily determined. Thus, the results generated by the disk diffusion are qualitative in contrast with dilution testing, in which the MIC is generated providing quantitative results. The advantages of the disk diffusion method are the low cost and the ease in modifying test formats when needed. However, because the susceptible category generated by the disk diffusion test shows reasonable but incomplete correlation with clinical outcome, there is a need for more precise interpretation of results from in vitro antibacterial susceptibility testing. 2 Studies comparing clinical response with in vitro data generated from dilution testing have shown that treatment failures are more frequently associated with higher MICs. 2,4 Although there are several methods for performing dilution antimicrobial susceptibility tests, the method most commonly used in veterinary diagnostic bacteriology laboratories is the broth microdilution test. Because most broth microdilution test panels used in veterinary medicine are prepared commercially, they are less flexible in adjusting to the changing needs of the practitioner and are more expensive to use than the disk diffusion method. However, the results generated from dilution 164
Use of computerized antimicrobial system with bacteria isolated from animals 16 testing allow the practitioner to alter the dose or frequency of administration based on the MIC of the pathogen. The BIOMIC a system was developed so that quantitative results could be generated from a testing method that had the flexibility and cost effectiveness of the disk diffusion test. The BIOMIC system provides both qualitative and quantitative information through the derivation of the MIC from the zone diameter measurement based on their inverse linear relationship. In other words, a large zone of inhibition will, in theory, correspond with a small MIC, and a small zone will correspond to a large MIC. This system is both reliable and cost effective when compared with traditional dilution testing on human bacterial pathogens. 1,3, However, its merits for use with animal pathogens have not been examined. In the present study, the MICs generated by the BIOMIC system were compared with those generated by the broth microdilution test for routinely used antimicrobial s tested against common bacterial pathogens of importance in veterinary medicine. 7 Materials and methods s. One hundred thirty-one veterinary clinical bacterial isolates were used for this study: Staphylococcus aureus ( isolates), Staphylococcus intermedius (), Enterococcus spp. (27), Pseudomonas aeruginosa (), Pasteurella haemolytica (29), Pasteurella multocida (26), Escherichia coli (9), and Klebsiella pneumoniae (). The reference strains used for quality control were S. aureus ATCC 29213 (microdilution tests), S. aureus ATCC 2923 (disk diffusion tests), and E. coli ATCC 2922. 7 Once isolated and identified, the clinical isolates were stored in sterile defibrinated sheep blood at 70 C. The day before testing, the organisms were thawed and streaked for isolation on a Columbia agar-based medium b supplemented with defibrinated sheep blood, c 1 yeast extract, d and 1 horse serum e (EBA). The agar plates were incubated for 24 hr at 3 37 C in an atmosphere of CO 2. Disk diffusion tests. Colonies of the test organisms that grew overnight were taken directly from the EBA plates and suspended in sterile water to an optical density equivalent to a 0. McFarland standard. Disk diffusion testing was performed according to the NCCLS M2-A6 document. 8 Mueller-Hinton (MH) agar b was used as the test medium for all organisms except, which was tested on MH agar supplemented with defibrinated sheep blood. The inoculated plates were stacked no more than 2 high and incubated for 24 hr at 3 37 C in ambient air. Zones of inhibition were read using an electronic caliper connected to a computer preprogrammed with the BIOMIC program. The BIOMIC system then calculated MICs from the agar diffusion gradient using regression line analysis. 3 Broth microdilution tests. Colonies of the test organisms that grew overnight were taken from the EBA plates and suspended in sterile water to an optical density equivalent to a 0. McFarland standard. Dilution tests were performed in accordance with the procedures outlined in the NCCLS M7- A4 document. 6 Broth microdilution trays were prepared in the laboratory. All antimicrobial s were prepared in accordance with the manufacturer s instructions or, when not available, with the NCCLS M7-A4 document guidelines. 6 Broth microdilution trays were stacked no more than 2 high and incubated for 24 hr at 3 37 C in ambient air. Interpretation of results. To compare MICs generated by the BIOMIC system with MICs generated by the broth microdilution method, the BIOMIC MIC was divided by the broth microdilution MIC (BIOMIC/MIC). The BIOMIC MICs were considered to be in with the broth microdilution MICs when the ratios were between 0. and 2.0. 3 The BIOMIC MICs were considered not in with the broth MICs when the ratios were 0.2 or 4.0. Because the MICs generated by the BIOMIC system were calculated to the nearest 0.1 g/ml, the ratios did not always match the doubling dilutions generated by the broth microdilution method. Thus, when the ratios fell between 0.2 and 0. they were expressed as 0., when they were between 0. and 1.9 the ratios were expressed as 1 and ratios between 2.0 and 3.9 were recorded as 2. If the MIC generated by both test methods was read either as less than or greater than, the ratio was considered to be 1. The ratio was also considered to be 1 if the BIOMIC MIC was less than and the broth microdilution MIC was less than or equal to or if the BIOMIC MIC was greater than and the broth microdilution MIC was greater than or equal to. Results Initially, the MIC between the BIOMIC and the broth microdilution method involved 231 bacteria antibacterial combinations. There was an 88 correlation between the 2 systems for the grampositive bacteria and an 8 correlation for the gramnegative bacteria (Table 1). The major discrepancy for the gram-positive bacteria was due to the Enterococcus spp., which showed only a 30 in MICs between the 2 methods when testing (Table 2). Although the overall for the gram-negative bacteria was, there were major discrepancies with and. There was a 9 between the 2 methods for P. haemolytica and a 78 for. When the results for individual bacteria drug combinations were examined, showed 70 for ampicillin and showed 70 for all 3 of the drugs tested. Because of these results, an additional 17 isolates of,, and were tested against ampicillin,, and plus additional antibacterial s: ciprofloxacin, trimethoprim/sulfamethoxazole, erythromycin, enrofloxacin, and tetracycline. The additional testing resulted in 432 bacteria drug combinations (Table 3). Agreement between the BIOMIC MICs and the MICs obtained from the broth microdilution method was 64. The BIOMIC MICs
166 Hubert, Nguyen, Walker Table 1. * obtained by BIOMIC and broth microdilution for all organisms tested. No. isolates Gram positive 30 ampicillin Gram negative 47 ampicillin 0.2 0. 1 2 4 2.1 4.3 23.3 3.3.6 2. 6.4 3.3 83.3 66 49 74.4 4.3.6 8. 26.7 17.6.6 73.3.9 8.1 89.3 for the were lower than the broth microdilution MICs for ampicillin, ciprofloxacin,, erythromycin, and enrofloxacin. Pasteurella multocida had lower BIOMIC MICs for and. However, all 3 organisms had BIOM- IC MICs that were greater than the MICs obtained from broth microdilution testing for trimethoprim/sulfamethoxazole, erythromycin, and tetracycline. Overall, there was a 76 for the isolates for bacteria drug combinations, 6 for the isolates, and 49 for the Cephalothin had an overall (0. 2) of with and. Ciprofloxacin and enrofloxacin both had an overall of with. However, only the ciprofloxacin combination resulted in a BIOMIC/MIC ratio of 1. Discussion The BIOMIC system is a reliable method for obtaining quantitative in vitro antimicrobial susceptibility results from the disk diffusion testing procedure when testing human bacterial pathogens. 1,3, However, some identification systems use databases that were generated using bacteria isolated from humans and thus are unreliable for identifying bacteria isolated from ani- Table 2. * obtained by BIOMIC and broth microdilution for individual organisms tested. No. isolates Staphylococcus aureus ampicillin Staphylococcus intermedius ampicillin ampicillin Pseudomonas aeruginosa ampicillin Pasteurella haemolytica 9 ampicillin Escherichia coli 9 ampicillin Klebsiella pneumoniae ampicillin Pasteurella multocida 9 ampicillin 0.2 0. 1 2 4. 60 70 60. 88.8 70 44.4 30. 99.9 99.9 88.8
Use of computerized antimicrobial system with bacteria isolated from animals 167 Table 3. * obtained by BIOMIC and broth microdilution for (17 isolates), Pasteurella multocida (17 isolates), and ( isolates). 0.2 0. 1 2 4 Ampicillin Cephalothin Ciprofloxacin Gentamicin Trimethoprim/sulfamethoxazole Eythromycin Tetracycline Enrofloxacin 8.8 2.9 29.4 94.1 3.3 8. 76. 29.4 82.4 41.2 7 2.9 2 3.3 2 4 1 4 2.9 4 1 94.1 2 41.2 82.4 8.0 47.0 76. 0 88.3.0 7.0 70. 0 8.0 76. 4.0.0 mals.,9 Thus, it was unclear how reliable the BIOMIC system would be for generating MIC data for bacterial pathogens from animals. The data presented here show that the BIOMIC system MICs correlate closely with the broth microdilution MICs for the majority of organisms tested against ampicillin,, and. For example, the for Pseudomonas aeruginosa was for ampicillin,, and. For E. coli, K. pneumoniae, S. aureus, and S. intermedius, the between the 2 methods was. However, there were discrepancies in MICs generated by the 2 systems when testing, Pasteurella multocida, and P. haemolytica. The trend with was toward lower BIOMIC MICs and therefore lower BIOMIC/. This trend could be attributed to a slow growth rate of Enterococcus on the MH agar medium. A slow growth rate would result in larger zones of inhibition, which would be interpreted by the BIOMIC as smaller MICs. Pasteurella haemolytica, however, demonstrated a trend towards higher BIOMIC MICs, possibly because of increased growth rate on the agar medium or decreased growth rate in the broth microdilution tray. If there were a reduced growth rate in the microdilution tray, it could have been due to incubating the trays aerobically rather than under CO 2. Overall there was no predictable trend for. Instead, the MIC seemed to be based more on the antibacterial being tested than the method tested. For example, against ampicillin, tended to have lower BIOMIC MICs, whereas when tested against erythromycin the BIOMIC MICs were greater than those generated by the broth microdilution method. The discrepancies between the BIOMIC MICs and the MICs generated by the broth microdilution method might be explained by the BIOMIC system s limited testing with veterinary isolates. Although Pasteurella spp. have been known to cause disease in humans, their involvement in infectious diseases remains primarily in veterinary medicine. The Pasteurella spp., therefore, may not have been adequately represented in the BIOMIC database. As seen previously, commercial systems that are useful for testing human pathogens need additional modifications to accurately test veterinary pathogens.,8 The data presented here suggest that the BIOMIC system is potentially useful to the veterinary community for obtaining quantitative results from in vitro antimicrobial susceptibility testing, but its database would have to be expanded to include veterinary pathogens. A larger study testing a wide variety of veterinary pathogens and antimicrobial
168 Hubert, Nguyen, Walker s would have to be performed to confirm the extent of the modifications needed by the BIOMIC system to service veterinary diagnostic laboratories. Sources and manufactures a. Giles Scientific, New York, NY. b. BBL, Becton-Dickinson, Cockeysville, MD. c. Cleveland Scientific, Bath, OH. d. Difco Laboratories, Detroit, MI. e. GIBCO Laboratories, Lawrence, MA. References 1. Berke I, Tierno PM Jr: 1996, Comparison of efficacy and costeffectiveness of BIOMIC VIDEO and Vitek antimicrobial susceptibility test systems for use in the clinical microbiology laboratory. J Clin Microbiol 34:19 1984. 2. Craig WA: 1993, Qualitative susceptibility tests versus quantitative MIC tests. Diagn Microbiol Infect Dis 16:231 236. 3. D Amato RF, Hochstein L, Vernaleo JR, et al.: 198, Evaluation of the BIOGRAM antimicrobial susceptibility test system. J Clin Microbiol 22:793 798. 4. Gerber AU, Craig WA: 1981, Worldwide clinical experience with cefoperazone. Drugs 22(Suppl.):8 118.. Matthews KR, Oliver SP, King SH: 19, Comparison of Vitek gram-positive identification system with API Staph-Trac system for species identification of staphylococci of bovine origin. J Clin Microbiol 28:1649 161. 6. National Committee on Clinical Laboratory Standards (NCCLS): 1993, Methods for dilution of antimicrobial susceptibility tests for bacteria that grow aerobically, 4th ed., approved standard M7-A4. NCCLS, Wayne, PA. 7. National Committee on Clinical Laboratory Standards (NCCLS): 1997, Performance standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals, tentative standard M31-T. NCCLS, Wayne, PA. 8. National Committee on Clinical Laboratory Standards (NCCLS): 1997, Performance standards for antimicrobial disk susceptibility tests, 6th ed., approved standard M2-A6. NCCLS, Wayne, PA. 9. Salmon SA, Watts JL, Walker RD, et al.: 199, Evaluation of a commercial system for the identification of gram-negative nonfermenting bacteria of veterinary importance. J Vet Diagn Invest 7:161 164.. Sautter RL, Denys GA: 1987, Comparison of BIOGRAM and commercial microdilution antimicrobial susceptibility test systems. J Clin Microbiol 2:301 304.