Antimicrobial Susceptibility Testing of Aquatic Bacteria: Quality Control Disk Diffusion Ranges for Escherichia coli ATCC 25922 and Aeromonas salmonicida subsp. salmonicida ATCC 33658 at 22 and 28{degrees}C

Quality control (QC) ranges for disk diffusion susceptibilitytesting of aquatic bacterial isolates were proposed as a resultof a multilaboratory study conducted according to proceduresestablished by the National Committee for Clinical LaboratoryStandards (NCCLS). Ranges were proposed for Escherichia coliATCC 25922 and Aeromonas salmonicida subsp. salmonicida ATCC33658 at 22 and 28°C for nine different antimicrobial agents(ampicillin, enrofloxacin, erythromycin, florfenicol, gentamicin,oxolinic acid, oxytetracycline, ormetoprim-sulfadimethoxine,and trimethoprim-sulfamethoxazole). All tests were conductedon standard Mueller-Hinton agar. With $>=$ 95% of alldata points fitting within the proposed QC ranges, the resultsfrom this study comply with NCCLS guidelines and have been acceptedby the NCCLS Subcommittee for Veterinary Antimicrobial SusceptibilityTesting. These QC guidelines will permit greater accuracy ininterpreting results and, for the first time, the ability toreliably compare susceptibility test data between aquatic animaldisease diagnostic laboratories.

INTRODUCTION

Top

Introduction

In 1998, the National Committee for Clinical Laboratory Standards(NCCLS) formed the Subcommittee on Veterinary AntimicrobialSusceptibility Testing-Aquaculture Working Group (VAST-AWG)to produce a guidance document for standardizing methods ofantimicrobial susceptibility testing (AST) of bacteria isolatedfrom aquatic animal species. The working group has since reliedheavily on the initial work of Barker and Kehoe (2), the effortsof those who organized the Workshop on MIC Methodologies inAquaculture (Weymouth, United Kingdom, 1998), and Alderman andSmith (1), who published the draft protocols developed at theworkshop. Alderman and Smith outlined the problems commonlyencountered when comparing data created by aquatic animal diseasediagnostic laboratories using different methods. The data generatedfrom these different methods differ greatly from laboratoryto laboratory, making it difficult to correlate susceptibilityresults between labs. The methods published by Alderman andSmith were termed "tentative" by the authors due to a numberof "unresolved issues" (1).

Building on the previous efforts, members of the present VAST-AWGhave targeted some of these unresolved issues, such as the developmentof quality control (QC) limits for potential QC strains forantimicrobial agents of interest in aquaculture. Some of theseantibiotics, though not yet approved for food source aquaculturepurposes, are prescribed for "extralabel" use by veterinarianstreating nonfood commercial and hobby aquarium fish. Since manyaquatic pathogens require lower incubation temperatures, theseorganisms cannot be tested accurately using the NCCLS AST methodsfor organisms whose optimal growth temperature is 35°C.

Although dilution susceptibility tests are becoming more popular,the most commonly used method of susceptibility testing in aquaticdiagnostic laboratories is still the disk diffusion method (2).Two QC strains were chosen for this multilaboratory disk diffusionstudy on the basis of their susceptibility profiles and theirinternational acceptance. Aeromonas salmonicida subsp. salmonicida(ATCC 33658 and NCIMB 1102) and Escherichia coli (ATCC 25922and NCIMB 12210) are both susceptible to a wide range of antimicrobials,grow well at low temperatures, and have been shown to be stablein the laboratory following multiple passes on artificial media.Aeromonas salmonicida subsp. salmonicida, in particular, waschosen because it has an optimum growth temperature lower thanthat of E. coli and is representative of aquatic pathogens.It was proposed to the NCCLS Subcommittee on VAST that bothof these organisms be used as QC organisms for disk diffusionsusceptibility testing of those aquatic isolates found in group1 (Table 1). Since there is a ban on importation of A. salmonicidasubsp. salmonicida strains in some nations, E. coli should beused instead in those nations.

View this table:
[in this window]
[in a new window]
TABLE 1. NCCLS VAST-AWG-recommended grouping for standardizing disk susceptibility tests of various bacteria isolated from fish

QC ranges were defined at two different temperatures, 22 and28°C, to accommodate optimum temperatures for aquatic bacteriaisolated from both warm-water and cold-water species. The specifictemperatures were chosen on the basis of routine use in aquaticanimal disease diagnostic laboratories worldwide, recommendationsof the AWG, and an effort to coordinate methodologies with internationalinvestigators.

The multilaboratory trial established QC ranges for nine differentantimicrobial agents: ampicillin, enrofloxacin, erythromycin,florfenicol, gentamicin, oxytetracycline, ormetoprim-sulfadimethoxine,trimethoprim-sulfamethoxazole, and oxolinic acid. These drugswere chosen to represent major classes of antimicrobial agents,some of which are approved for use in the United States and/orabroad. In addition, some of these drugs have been identifiedin the environment (3, 4) and are of growing concern to environmentalregulatory agencies. Results from this study will assist inthe accurate monitoring of resistance in bacteria commonly isolatedfrom the environment.

MATERIALS AND METHODS

Top

Materials and Methods

Participating laboratories. In this study, data were generated in 10 participating laboratories.These included the Animal Health Diagnostic Laboratory, MarylandDepartment of Agriculture, College Park, Md.; Idaho Fish HealthCenter, U.S. Fish and Wildlife Service, Ahsahka, Idaho; AtlanticVeterinary College, University of Prince Edward Island, PrinceEdward Island, Canada; School of Veterinary Medicine, LouisianaState University, Baton Rouge, La.; Virginia Department of Agricultureand Consumer Services, Warrenton, Va.; Animal Health Division,Alpharma, Chicago Heights, Ill.; Wisconsin Veterinary DiagnosticLaboratory, University of Wisconsin, Madison, Wis.; WashingtonAnimal Disease Diagnostic Laboratory, Washington State University,Pullman, Wash.; Laboratory of Public Health, University of Patras,Rio Patras, Greece; and the Center for Veterinary Medicine,Division of Animal and Food Microbiology, Food and Drug Administration,Laurel, Md.

Despite the fact that this study was initiated with 10 participatinglaboratories, QC ranges were chosen on the basis of data fromnine testing sites. One laboratory's values were consistentlylower than those generated in the other participating laboratoriesdue to the use of a magnifying lens rather than the unaidedeye to interpret the zones of inhibition. As a result, the totalQC data points per organism per drug per temperature per incubationtime condition were reduced from 600 to 540.

Antimicrobial compounds. All antimicrobial disks were obtained from BD Diagnostic Systems(Sparks, Md.). The nine antimicrobials tested and their correspondingdisk quantities and disk lot numbers were as follows: ampicillin,10 µg (lot no. 1179720 and 1277720); enrofloxacin, 5 µg(1346726 and 1332725); erythromycin, 15 µg (1267734 and1297736); florfenicol, 30 µg (1052735 and 1129733); gentamicin,10 µg (1191729 and 1276721); oxytetracycline, 30 µg(0054734 and 1309724); ormetoprim-sulfadimethoxine, 1.25 and23.75 µg (1254725 and 1317725); trimethoprim-sulfamethoxazole,1.25 and 23.75 µg (1267736 and 1306720); and oxolinicacid, 2 µg (1302722). In this study, only one lot of oxolinicacid disks was available to all laboratories, but a limitedsupply of another lot (1046728) was made available to one ofthe participating laboratories.

Test strains and growth conditions. E. coli reference strain ATCC 25922 and A. salmonicida subsp.salmonicida reference strain ATCC 33658 were used to establishQC ranges at both 22 ± 2 and 28 ± 2°C. TheCenter for Veterinary Medicine at the Food and Drug Administrationconducted an internal QC study with E. coli ATCC 25922 (usingenrofloxacin, ampicillin, gentamicin, trimethoprim-sulfamethoxazole,oxytetracycline, and florfenicol), Pseudomonas aeruginosa ATCC27853 (using gentamicin), and Enterococcus faecalis ATCC 29212(using trimethoprim-sulfamethoxazole) as reference strains at35°C with the three lots of Mueller-Hinton (MH) agar usedin this study. All QC testing data points obtained from thereference strains at 35°C fell within the established QClimits found in NCCLS document M31-A2 (13).

Disk diffusion susceptibility testing. The study was designed in accordance with NCCLS guidelines M23-A2(11) and M37-A2 (12). On 10 testing days, each laboratory inoculatedeach of the control strains onto 12 MH agar plates. The MH agarwas prepared by Prepared Media Laboratory (Wilsonville, Oreg.),as described in the NCCLS document M31-A2 (13) and distributedto the participating laboratories. Three different MH agar lotswere used: Acumedia (Baltimore, Md.) catalog no. 0109-126, Difco(Sparks, Md.) catalog no. 112185-1, and BBL (Baltimore, Md.)catalog no. 112372-1. The nine antimicrobial disks were appliedto the media. Zones of inhibition were read after 24 to 28 hand 44 to 48 h (22°C) and after 24 to 28 h (28°C). Laboratoriestested the strains in parallel on each of the 10 test days underthe conditions indicated above. Four independent suspensionswere taken from four separate culture plates (E. coli [22 and28°C] and A. salmonicida subsp. salmonicida [22 and 28°C]),resulting in a total of 540 QC data points (1 suspension x 3lots of media x 2 lots of disks x 10 test days x 9 laboratories)for each condition.

Proposed QC strains were each tested 20 times with each agarlot for all 10 antimicrobial agents assayed. This resulted ina total of 60 possible tests by each laboratory for each antimicrobialagent and a target of 540 test values for each organism perdrug per temperature per incubation time condition. For oxolinicacid, 297 total test values were recorded (30 tests per laboratoryplus an additional 27 tests performed by one laboratory whichhad a limited number of disks from another lot).

Testing protocol. The study was performed according to the methods described inNCCLS report M42-R (9). Using a Sensi-disk self-tamping 12-placedispenser (BBL), nine antimicrobial disks were applied to each15- by 150-mm MH agar plate. Plates were stacked no more thanfour high and placed in ambient air incubators at 22 or 28°C.

Definition of zones of inhibition. Each area that was detected with the unaided eye as showingno obvious, visible growth was recorded as a zone of inhibition.Faint growth or tiny colonies that were detected only with difficultyat the edge of a zone of inhibited growth were not considered.When the potentiated sulfonamides were tested, antagonists inthe medium allowed some growth (6); therefore, slight growthwas disregarded and the margin of heavy growth (>80% of alawn) was used to determine the zone diameter.

RESULTS AND DISCUSSION

Top

Results and Discussion

QC ranges were developed at 22°C (24 to 28 h and 44 to 48h) and 28°C (24 to 28 h) for E. coli ATCC 25922 and A. salmonicidasubsp. salmonicida ATCC 33658. The organisms were tested 540times per drug per temperature condition. In accordance withNCCLS guidelines M23-A2 (11) and M37-A2 (12), the percentageof participant zone diameters that fell within the approvedQC ranges exceeded 95% for each antibacterial agent tested.

MH agar was selected as the growth medium on the basis of itsability to support the growth of many aquatic isolates as wellas of the QC strains. It is also specified for aerobic disksusceptibility testing methods (10) and is ion adjusted andtherefore standardized. Additionally, this medium has showngood intralaboratory reproducibility of susceptibility resultsof tests using the QC strains as well as various aquatic isolates.The three lots of MH agar used in this study yielded minimallot-to-lot variation within and between the laboratories forboth QC organisms tested. Similarly, very little variation wasobserved in the zones of inhibition measured around the twodifferent lots of antimicrobial disks tested.

Table 2, Table 3, and Table 4 summarize the zones of inhibitionand QC limits for the nine antimicrobial agents tested for E.coli. The NCCLS-approved zone diameter QC ranges for the organismper drug per temperature per incubation time condition weredetermined by using a modification of the median method describedby Gavan et al. (5). Where appropriate, the calculated rangeswere adjusted to incorporate more or fewer of the data pointsas long as $>=$ 95% were included.

View this table:
[in this window]
[in a new window]
TABLE 2. Disk diffusion QC results for E. coli ATCC 25922 at 22°C and 24 to 28 h with MH agar^a

View this table:
[in this window]
[in a new window]
TABLE 3. Disk diffusion QC results for E. coli ATCC 25922 at 22°C and 44 to 48 h with MH agar^a

View this table:
[in this window]
[in a new window]
TABLE 4. Disk diffusion QC results for E. coli ATCC 25922 at 28°C and 24 to 28 h with MH agar^a

When tested with enrofloxacin, E. coli yielded ranges considerablywider than those presently being used in disk diffusion testingof veterinary isolates at 35°C (13). To strengthen the testingmethod, all ranges for E. coli tested with enrofloxacin weredeemed unacceptable. When E. coli was tested with ampicillin,distinct inner and outer margins, some resembling a halo orring of growth forming the inner margin, were observed. Somezones had individual resistant colonies inside the inner zoneof inhibition. The ranges, however, were narrow and consistentamong most of the laboratories, thus permitting ranges to beestablished for ampicillin.

Aeromonas salmonicida subsp. salmonicida produced zones of inhibitionwith NCCLS-approved ranges of $<=$ 16 mm (Table 5, Table 6, and Table7), which were considered acceptable QC ranges under the testconditions for all antimicrobial agents tested. When testedagainst the potentiated sulfonamides, A. salmonicida subsp.salmonicida yielded a very distinct inner and outer margin ofgrowth. The antagonists in the medium allowed some slight growthinside the outer margin; however, this was not observed withE. coli. Measuring zones around these disks with A. salmonicidasubsp. salmonicida as a QC organism may result in some laboratory-to-laboratoryvariation in the interpretation of the margin, so E. coli maybe a more suitable QC organism when testing the potentiatedsulfonamides. When tested with all drugs except gentamicin andenrofloxacin, larger zones of inhibition were observed at both22 and 28°C with A. salmonicida subsp. salmonicida thanwith E. coli. Despite being active against primarily gram-positiveand rapidly dividing bacteria, erythromycin was shown to beactive against both of the E. coli and A. salmonicida subsp.salmonicida strains under in vitro conditions at the temperaturesused in this study.

View this table:
[in this window]
[in a new window]
TABLE 5. Disk diffusion QC results for A. salmonicida subsp. salmonicida ATCC 33658 at 22°C and 24 to 28 h with MH agar^a

View this table:
[in this window]
[in a new window]
TABLE 6. Disk diffusion QC results for A. salmonicida subsp. salmonicida ATCC 33658 at 22°C and 44 to 48 h with MH agar^a

View this table:
[in this window]
[in a new window]
TABLE 7. Disk diffusion QC results for A. salmonicida subsp. salmonicida ATCC 33658 at 28°C and 24 to 28 h with MH agar^a

Results at the lower temperatures used here showed a clear decreasein the precision of the disk diffusion test, with wider rangesresulting for some of the antimicrobials tested compared tothose ranges presently used at 35°C (14). The effect oftemperature on the intralaboratory and approved QC ranges ofmost of the test articles was apparent, with narrower rangesresulting when testing at 28 versus 22°C after 24 h. Thisrange narrowing indicates that at 28°C, margins were moredefined than at 22°C, probably due to an increased growthrate or to a temperature-dependent decrease in the drug diffusionrate at the lower temperature. Wider ranges were observed inmost cases, especially with A. salmonicida subsp. salmonicida,at 22°C after 48 h versus those measured after 24 h at 22and 28°C. This suggests that the precision of the methoddecreased slightly over time.

It is recognized that some of these approved ranges may be widerthan those used for the same drugs at higher temperatures, butthe investigators believe this is a direct effect of the decreasedincubation temperature. The growth kinetics of these two organismsat the temperatures used for testing was not determined butmay be a subject of interest in the future.

The methods used and ranges proposed were presented to the NCCLSSubcommittee on VAST. All methods and ranges have been acceptedfor inclusion in NCCLS report M42-R (9).

This is the first large-scale study to establish QC ranges forAST of aquatic isolates at temperatures lower than the standard35°C. It is envisioned that these ranges can be a nidusfor future studies to establish ranges for more drugs usingthis core method and to develop additional standardized testingmethods for fastidious aquatic pathogens (Table 1).

ACKNOWLEDGMENTS

We extend our sincere thanks to the support staff of the participatinglaboratories, especially David Melka, Gina Cheuk, Lorraine Lund,Doris Poole, and Marilyn Blair. We also thank Danae Venierifor her hard work during the study.

This study was made possible by financial support from Schering-PloughInc. (Kenilworth, N.J.), BD Diagnostic Systems (Sparks, Md.),and the Oak Ridge Associated Universities sponsored by the Foodand Drug Administration.

FOOTNOTES

* Corresponding author. Mailing address: U.S. Food and Drug Administration, Center for Veterinary Medicine, HFV530, 8401 Muirkirk Rd., Laurel, MD 20708. Phone: (301) 827-7902. Fax: (301) 827-8250. E-mail: RMiller1{at}cvm.fda.gov.

REFERENCES

Top