Antibiotic Susceptibility of Anaerobic Bacteria

Transcription

1 AwNrIcRoBuL AIGENTS AND CHUMOT R"PY, Mar. 1973, p. 35O-356 Copyright 1973 American Society for Microbiology Vol. S, No. 3 Prind in U.S.A. Modified Broth-Disk Method for Testing the Antibiotic Susceptibility of Anaerobic Bacteria TRACY D. WILKINS AND TERESA THIEL Anaerobe Laboratory, Division of Basic &iences, College of Agriculture, Virginia Polytechnic Institute and State University, Blackburg, Virginia Received for publication 2 October 1972 The most commonly used method for testing the antibiotic susceptibility of aerobic and facultative bacteria is the disk diffusion method. However, some anaerobic bacteria do not grow well enough in anaerobic jars for performance of disk diffusion tests. A modification of the broth-disk method of Schneierson allowed us to determine antibiotic susceptibility in a completely anaerobic environment. Commercial antibiotic disks were added anaerobically to tubes of prereduced brain heart infusion broth to achieve a concentration of each antibiotic approximating that attainable in blood. The tubes were then inoculated and incubated for 18 h. Resistance or susceptibility to each antibiotic was determined according to the amount of growth in each tube as compared with a control culture without the antibiotic. There was good correlation between results obtained by this broth-disk method and minimal inhibitory concentrations. Anaerobic bacteria have been implicated as the causative agents in several types of tissue infections, and clinical laboratories are often requested to process samples for anaerobes. Disk diffusion methods for testing the antibiotic susceptibility of anaerobic bacteria have been described (1, 8); however, until recently, no standardized methods based on interpretative zone diameters were available. Two disk diffusion methods have been standardized for several antibiotics by the use of regression plots of ininimal inhibitory concentrations (MIC) versus zone diameters (10, 11). However, these tests are limited to use with common, rapidly growing anaerobic isolates which are not extremely oxygen-sensitive. During a small-scale clinical trial of the disk diffusion method of Wilkins et al. (11), we recognized that a simple method was needed which could be used to determine the susceptibility of both slow-growing and very oxygen-sensitive organisms that could not be tested by this disk diffusion technique. The principles of Schneierson's method (9) for use with facultative bacteria have recently received increased usage. In this method, which we have termed "broth-disk," a concentration of an antibiotic approximately equal to that achievable in blood is put into a tube of broth medium with commercial filter-paper disks as the carrier of the antibiotic. Growth of the organism at that concentration is considered to constitute resistance to clinically useful levels of the antibiotic. This method simply designates whether the MIC is above or below concentrations normally attained in the blood. Abramson and Smibert (Brit. J. Vener. Dis., in press) recently reported the use of a modified version of the broth-disk method for assaying the antibiotic susceptibility of spirochetes, which cannot be tested by any disk diffusion method. The need for automation in antibiotic susceptibility testing also has resulted in revival of this idea. Isenberg and colleagues have described, in a series of publications (5-7), an automated machine for testing the antibiotic susceptibility of facultative bacteria. This instrument uses antibiotic disks to carry a single concentration of an antibiotic into a vial of broth which is then inoculated with a standard inoculum. The results obtained from this instrument have reportedly been as accurate as results achieved by the standard Kirby-Bauer (6) disk diffusion method. These authors have shown that the antibiotic completely diffuses from the disk in liquid media within 1 min (5, 7). The present paper describes the use of Schneierson's original idea as a simple system for testing the antibiotic susceptibility of anaerobic bacteria. The results obtained with clinically important anaerobic bacteria are compared with results of MIC determinations and with results obtained by the disk diffusion method of Wilkins et al. (11). MATERIALS AND METHODS Bacterial strains. The strains of anaerobes used in this study were from the culture collection 350

2 VOL. 3, 1973 BROTH-DISK TEST FOR ANAEROBIC BACTERIA 351 of the Anaerobe Laboratory. The majority were clinical isolates sent to this laboratory for identification. Organisms were identified by methods previously described (3). Media. Prereduced media were prepared according to published methods and formulas (3) and contained 0.05% cysteine. The medium used for antibiotic susceptibility testing was prereduced BBL brain-heart infusion broth, supplemented with % hemin, 0.002% menadione, and 0.5% yeast extract (BHI-S). Inocula were grown in prereduced chopped meat (CM), chopped-meat carbohydrate (CMC), peptone-yeast extract-glucose (PYG), BHI-S, or Thioglycollate (Difco) broth. Antibiotics. Sources for antibiotic powders were as follows: penicillin G, ampicillin, and tetracycline (E. R. Squibb and Sons, Inc., New Brunswick, N.J.); clindamycin (The Upjohn Co., Kalamazoo, Mich.); erythromycin (Abbott Laboratories, Chicago, Ill.); and chloramphenicol (Parke, Davis & Co., Detroit, Mich.). Erythromycin and chloramphenicol were dissolved in ethanol; ampicillin was dissolved in sterile, ph 8.0 phosphate buffer (0.1 M). All other antibiotics were dissolved in sterile distilled water. Determination of MIC. All MIC determinations were made by the broth dilution method previously published (11). The MIC determinations with erythromycin were performed under oxygen-free nitrogen instead of carbon dioxide, which was used for the other antimicrobial agents. Disk diffusion method. Results of a single determination by the disk diffusion method (11) were used for calculations of correlation with MIC values. For purposes of these tests, intermediate zone diameters were reported as resistant. Broth-disk susceptibility determinations. Rubber-stoppered tubes (18 by 142 mm), each containing 5 ml of BHI-S medium, were opened anaerobically under 0-free CO2 or N2, and the appropriate number (Table 1) of antibiotic disks were added aseptically from a commercial cartridge (Difco). The tubes were simultaneously inoculated from a Pasteur pipette with one drop of an 18- to 24-h culture of the test organism in CM broth. A control culture containing no antibiotics was included in each set. Incubation was at 37 C for 18 to 24 h. Susceptibility to the test antibiotic was defined as either absence of turbidity or less than 50% of the turbidity of the control culture. In some cases, the inoculum was obtained from 24-h cultures in PYG, BHI-S, CMC, or Thioglycollate broth. Results of a single determination were used for calculations of the correlation of the method with MICs. Antibiotic disks. All antibiotic disks used for the broth-disk and disk diffusion tests were purchased from Difco and were the standard highconcentration disks. Disks were stored refrigerated in bottles containing desiccant. In one experiment, antibiotic disks obtained from Pfizer Co., Inc., and BBL were used. RESULTS AND DISCUSSION Inoculum size and source. The same size and source of inocula were used for both the broth-disk and MIC methods. One drop from a Pasteur pipette of an 18- to 24-h culture of the test organism was added to each tube containing 5 ml of BHI-S broth to give approximately a 1:100 dilution of the culture. This was an inoculum of 106 to 107 bacteria per ml, which is 10- fold higher than the inoculum usually used for broth MIC determinations with facultative bacteria (2). We chose to use the higher inoculum because some anaerobic bacteria did not grow as well from smaller inocula and sometimes required more than 24 h of incubation to reach maximal turbidity in the BHI-S broth. As a source of inoculum, we routinely used an 18- to 24-h culture in CM broth because, in our experience, this medium gave very reproducible growth of anaerobic bacteria. CM broth is also more difficult to oxidize than other anaerobic media. We found that other media could be used as sources of inocula without significantly changing the results of either the broth-disk method or the MIC determinations. When 10 strains of Bacteroidesfragilis, 5 strains of Peptostreptococcus anaerobius, and 5 strains of Fusobacterium varium were tested by the broth-disk method for susceptibility to five antibiotics with the use of inocula grown in PYG, BHI-S, CMC, and Thioglycollate broth, the results obtained with the four sources of inocula were the same as the CM results in 390 of the total 400 tubes. MIC determinations with penicillin gave the same results with all five sources of inocula when two strains of Peptococcus magnus and one strain each of Bacteroides fragilis, Peptostreptococcus anaerobius, Fusobacterium necrophorum, Clostridium perfringens, Peptostreptococcus intermedius, and Bacterioides clostridiiformis were tested. A light background turbidity was present in many of the antibiotic tubes that were inoculated with CMC broth because of the higher inocula obtained from this carbohydrate-containing medium. This could be eliminated by the use of a 1 :10 dilution of the CMC broth culture as an inoculum. Reading results. With the CM inoculum, interpretation of results in approximately 95% of the tests involved merely a judgment of growth versus no growth. Organisms that were resistant to an antibiotic normally grew to almost the same turbidity as the control culture, whereas susceptible organisms did not grow in the tubes containing that antibiotic. With a few organisms, the number of cells in the inoculum was high enough so that the initial generations that occurred prior to inhibition by antibiotics resulted

3 352 WILKINS AND THIEL in a light turbidity that was present with all of the antibiotics tested. In such cases, this background turbidity was disregarded, and only those tubes in which the turbidity was 50% or greater than the control culture were considered resistant. Incubation time. We recorded both the broth-disk and MIC results after incubation for 18 to 24 h. Resistance to an antibiotic could sometimes be reported much earlier, however, since both the control cultures and some cultures containing antibiotics were often very turbid after only 4 h of incubation. Definite susceptibility could not be reported until after 18 h of incubation, since growth could occur later in the tubes containing other antibiotics. However, presumptive susceptibility could be reported very early and confirmed after 18 h of incubation. Antibiotic concentrations. Opinions of various investigators differ as to the concentration of each antibiotic to which an organism must be resistant in order to be termed "clinically resistant." Listed in Table 1 are the test concentrations we chose for this trial of the brothdisk method. Except for chloramphenicol and erythromycin, the test concentrations were the same as the MIC break-points used in the disk diffusion method of Wilkins et al. (11). Other concentrations could be used by changing either the number of antibiotic disks or the amount of broth used per tube, and more than one concentration could be tested for each antibiotic. Correlation of the results of the brothdisk, disk diffusion, and MIC methods. To determine how well the results of the broth-disk and the disk diffusion methods correlated with MIC values, we compared the three methods using the 145 strains of anaerobic bacteria listed in Table 2. As indicated in Table 2, certain of the species were chosen for their significance in clinical infections, others for their poor growth characteristics which would test the methods most severely, and some species for both of these characteristics. Two of the more common clinical TABLE 1. Test concentrations of antibiotics Labeled No. of Calculated Antibiotic disk disks test conen content tuer per ml Penicillin G.. 10 units 1 2 units Ampicillin pg 2 4 pg Cephalothin... 30,g 1 6 pg Tetracycline pg 1 6 pg Clindamycin...2 pg pg Chloramphenicol.. 30 jpg 2 12 pg Erythromycin pg 1 3,ug ANTIMICROB. AG. CHEMOTHER. TABLE 2. Organisms used and reason they were chosen Charae- No. of Common teristic Species strains clinical poor tested isolates growth on plates Bacteroides fragilis Peptococcus magnus Peptostreptococcus anaerobius P. intermedius Eubacterium lentum Clostridium perfringens 5 + Bacteroides melaninogenicus Fusobacterium varium F. mortiferum F. russii Bacteroides ruminicolạ B. putredinis Actinomyces israelii 2 + A. viscosus A. naeslundii isolates, B. fraghls and P. magnus, were chosen for detailed analysis, and 50 strains of each of these species were tested. P. magnus grows only to a very low turbidity in broth, and on agar plates the growth is barely visible so that zones of inhibition are difficult to see. B. fragilis is a very fast-growing organism which reaches high turbidity in both broth and agar media. We also tested a total of 45 strains belonging to 11 other species of anaerobic bacteria to determine whether the broth-disk method could be used for a wide range of anaerobes. The susceptibility of each of the 145 strains to the seven antibiotics listed in Table 1 was determined by both the broth-disk and MIC methods. The MICs for the organisms tested were within two double dilutions of the broth-disk concentration in 26% of the tests. Listed in Table 2 are the number of tests in which the broth-disk results did not correlate with the MIC data and the calculated percentage accuracy of the broth-disk method with each of the species tested. The overall correlation of the broth-disk results with MIC values was 97% in these 1,015 tests. The results of the disk diffusion and MIC methods were also compared, but susceptibility to chloramphenicol and erythromycin was not tested by the disk diffusion method since interpretative zone diameters for these antibiotics are not available. The number of cases in which the results of the disk diffusion assays did not

4 VOL. 3t 1973 BROTH-DISK TEST FOR ANAEROBIC BACTERIA 353 TABLE 3. Accuracy of the broth-disk and disk diffusion tests as compared with MIC results Broth-Disk test Disk diffusion test Species tested No. of No. in- Percent No. of No. in- Percent testsa correctb correct testse correctb correct Bacteroides fragilis Peptococcus magnus c Peptostreptococcus anaerobius P. intermedius Eubacterium lentum c 3 88 Ciostridium perfringens Bacteroides melaninogenicus Fusobacterium varium F. mortiferum F. russii c Bacteroides ruminicola B. putredinis c Actinomyces israelii c a A. viscosuṡ A. naeslundii Total of all species... 1, a A test was defined as one strain tested for susceptibility to one antibiotic a single time. b Incorrect results were those which did not correlate with the MIC value. c One or more strains could not be tested by the disk diffusion method. d A dash means no data available. correlate with the MIC results is given in Table 3, as is the correlation of the results of the disk diffusion method as compared with the MIC values. With these species, the disk diffusion method correlated with MIC values in 92% of the tests. However, the disk diffusion method was only 8% correct in tests with the five strains of Clostridium perfringens. For this organism, the zone diameters were slightly smaller than the published break-points for susceptibility; yet the MIC results indicated that the five strains were uniformly susceptible to these antibiotics. Thus, it appears that smaller zone diameter breakpoints should be considered for the disk diffusion test with this organism. The disk diffusion method could not be used for testing the antibiotic susceptibility of eight other strains (three of P. magnus, two of Actinomyces israelii, and one each of Eubacterium lentum, Fusobacterium russii, and Bacteroides putredinis), since these organisms would not grow well enough in the agar medium for zones to be discerned. Many other strains of these same species grew poorly on the susceptibility plates, but zones could be measured, albeit with some difficulty. The percentage of organisms classified as susceptible to the antibiotics by each of the three methods is given in Table 4. The major discrepancy was that 88% of the strains of P. magnus were susceptible to erythromycin as determined by the MIC method, but only 44% appeared to be susceptible by the broth-disk method. Since it has been reported that erythromycin is less active at the lower ph produced by a CO2 atmosphere (4) the MIC determinations with erythromycin were performed with N2 as the oxygen-free gas. The broth-disk determinations were routinely performed with CO2 since this is the gas most widely used with prereduced media. When the broth-disk method was repeated with N2, there were no significant differences in the results (Table 4). We cannot explain this discrepancy between MIC and broth-disk results which occurred only in the erythromycin tests with P. magnus and not with the other 12 species. However, the overall correlation of MIC values with broth-disk tests of P. magnus with all seven antibiotics was 93%. Possible sources of error. The most obvious source of error in the broth-disk method is the actual concentration of antibiotic contained in commercial antibiotic disks. Since the Food and Drug Administration (FDA) has enforced standards on antibiotic disk production in the United States, the concentration in commercial disks is now closer to the labeled value. The FDA routinely determines a mean value for six disks from each batch of antibiotic disks manufactured in the United States. This mean value must be

5 354 WILKINS AND THIEL ANTIMICROB. AG. CHEMOTHER. TABLE 4. Percentage susceptibility of organisms as determined by three methods Antibiotics B. fragili, 50 strains P. magnus, 50 strains Other species, 45 strains Disk MIC Broth- Disk MIC Broth- Disk MIC Broth- diffudisk diffusion disk diffusion disk so Penicillin G Ampicillin Cephalothin Clindamycin Tetracycline Chloramphenicol Erythromycin (CO2) Erythromycin (N2)b a Tests performed with 02-free C02. b Tests performed with 02-free N2. TABLE 5. Information supplied by the Food and Drug Administration on actual contents of batches of accepted disks over a 12-month period Antibiotic disk Labeled concn Mean Manlbacnceo conce of nrage of all batches deviation Standard Ampicillin pg Cephalothin... 30gg Chloramphenicol g Clindamycin... 2 ;g Erythromycin s;g PenicillinG. 10 units Tetracycline ug between 67 and 150% of the labeled value or the batch of disks cannot be sold. Data on all assays performed by the FDA on the types of antibiotic disks used in the present study were supplied to us by Richard Norton and Alfred Giovetti (Table 5). The composite means for all of the batches of each type of disk were very close to the labeled values. The maximal variation between batches was 73 to 148% of the labeled value. Differences of this magnitude could influence the results of the broth-disk test but, since very few organisms have MICs within a twofold dilution of the break-point, the discrepancies are not numerous. In the process of obtaining the results reported here, several batches of antibiotics were used, and there was no evident difference between results obtained with different batches. Antibiotic disks are a convenient source of antibiotics for testing purposes, but the disks must be handled with care to prevent deterioration of the antibiotic. Antibiotic disks are most stable when stored at 4 C, or preferably at -20 C, in sealed bottles containing a desiccant. The precautions suggested by the International Collaborative Study on Antibiotic Susceptibility Testing (2) should be followed for both storage and handling of antibiotic disks, and all disks should be discarded by the expiration date. Contamination has not been a problem in our laboratory in the performance of the broth-disk tests; however, this could be a problem if the disks were heavily contaminated with bacteria. Of several thousand tubes used in our trials of the broth-disk test, we have found obvious contamination in approximately 10 tubes. Currently, in the United States, disks are manufactured chiefly by Difco, BBL, and Pfizer Co., Inc. We tested 100 disks at random from each manufacturer for levels of contamination that would interfere with the broth-disk test. Each disk was incubated for 48 h anaerobically in 5 ml of BHI-S broth and no turbidity developed in any tube. Erroneous results due to introduction of resistant bacteria on the antibiotic disks would thus seem to be very unlikely. However, any antibiotic tube in which the type of growth appears significantly different from the growth in the control tube should be stained to test for possible contamination.

6 VOL. 3, 1973 Organisms for which the MIC is very close to the test concentration may be reported as resistant in one broth-disk assay and susceptible in a later assay. This kind of variation is also inherent in MIC procedures in which a difference of a twofold dilution is considered normal variation between assays (2). Organisms with an MIC slightly below the test concentration also may appear resistant if the results are not recorded until well after 24 h of incubation. Prolonged incubation results in a gradual increase in the MIC of most antibiotics. Inoculum size does not have a great effect on either the broth-disk or MIC determinations with most anaerobic bacteria. We have found that a 10-fold dilution of the inoculum resulted in lowering the MIC by, at most, a twofold dilution. With some anaerobes, such as B. melaninogenicus the use of dilutions of inocula which are greater than 1:10 can result in erratic growth. In our laboratory it is much more convenient to use larger inocula and disregard the infrequent "background" turbidity that sometimes occurs, instead of making an anaerobic dilution to use as an inoculum. Other laboratories may prefer to use a 1:10 dilution in order always to have a growth versus no growth assay. Advantages of the broth-disk method. The principal advantage of the broth-disk method is that almost all clinical isolates can be tested for antibiotic susceptibility by this one procedure. Although in this trial we reported results for the diffusion test for all except eight of the strains tested, growth was so poor with many of the slow-growing anaerobes that zones of inhibition on agar plates were very difficult to discern. The results of the broth-disk method were easier to read and there was less variation in the results recorded for the same test by different technicians. The percentage of anaerobic isolates which cannot be tested by the disk diffusion method would vary among laboratories according to the types of specimens examined and the isolation techniques used. Of the last 100 anaerobic bacteria submitted to this laboratory by research and clinical laboratories for identification, 16 did not grow in the pour plates used for the disk diffusion test, and an additional 8 did not grow well enough for accurate measurements of zone diameters. Of these same 100 organisms, only 3 did not grow well enough in the prereduced broth for performance of the broth-disk test. Thus, the broth-disk test increased by 21% the number of organisms on which we could report antibiotic susceptibility data. In cases of endocarditis and certain other infections, the bactericidal activity of antibiotics is of much more importance than bacteriostatic BROTH-DISK TEST FOR ANAEROBIC BACTERIA 355 effects. Neither disk diffusion tests nor agar dilution MIC determinations can yield such information. The preferred method is to perform a broth dilution MIC determination with subsequent subculture of tubes at and above the MIC concentration to test for viable cells. Just as the broth-disk method can be used to approximate whether an organism can grow at concentrations of antibiotics attained in the blood, tubes containing antibiotics that inhibit growth in the broth-disk test can be subcultured to determine whether this antibiotic concentration was also bactericidal. One disadvantage of the broth-disk method is that it does not distinguish between degrees of resistance or susceptibility. Such data theoretically could be obtained for disk diffusion tests by consulting regression plots of MICs versus zone diameters, but in clinical practice this is not done routinely. The main disadvantages of the broth-disk method are cost of the tubes of prereduced BHI-S media and the requirement for a supply of 02-free CO2 or N2. The broth-disk method would seem particularly useful for clinical laboratories which currently identify anaerobes according to methods described in the Anaerobe Laboratory Manual (3), since susceptibility tests could be performed at the same time as the biochemical tests by the same technician. However, the method should work the same if performed inside an anaerobic glove box with anaerobic media. The choice of methods would depend on the requirements of individual laboratories. ACKNOWLEDGMENTS This project was supported by Public Health Service grant from the National Institute of General Medical Sciences and grant FR07095 from the National Institutes of Health, Biomedical Sciences. LITERATURE CITED 1. B6erens, H., and J. Guillaume Technique de determination de la sensibilit6 des bact6ries ana6robies aux antibiotiques par la m6thode des disques. Ann. Inst. Pasteur (Paris) 86: Ericsson, H. M., and J. C. Serris Antibiotic sensitivity testing. Report of an international collaborative study. Acta Pathol. Microbiol. Scand., Sect. B (Suppl. 217): Holdeman, L. V., and W. E. C. Moore (ed.) Anaerobe laboratory manual. Virginia Polytechnic Institute Anaerobe Laboratory, Blacksburg. 4. Ingham, J. B., J. B. Selkon, and J. H. Hale The effect of carbon dioxide on the sensitivity of Bacteroids fragili to certain antibiotics in vitro. J. Clin. Pathol. 23: Isenberg, H. D Development of an automated method to determine antibiotic susceptibility of rapidly growing organisms. Advan. Automated Anal. 1:

7 356 WILKINS AND THIEL 6. Isenberg, H. D., A. Reichler, and D. Wiseman Prototype of a fully automated device for determination of bacterial antibiotic susceptibility in the clinical laboratory. Appl. Microbiol. 22: Isenberg, H. D., and J. Seo Experimental foundation for the automated antibiotic susceptibility test. Advan. Automated Anal. 1: Merritt, E. S A simple method for the determination of the antibiotic-sensitivity of anaerobic organisms. Amer. J. Clin. Pathol. 38: ANTIMICROB. AG. CHEMOTHER. 9. Schneiersen, S. S A simple rapid disc-tube method for determination of bacterial sensitivity to antibiotics. Antibiot. Chemother. 4: Sutter, V. L., Y. Y. Kwok, and S. M. Finegold Standardized antimicrobial disc susceptibility testing of anaerobic bacteria. I. Susceptibility of Bacteroides fragili to tetracycline. Appl. Microbiol. 23: Wilkins, T. D., L. V. Holdeman, I. J. Abramson, and W. E. C. Moore Standardized singledisc method for antibiotic susceptibility testing of anaerobic bacteria. Antimicrob. Ag. Chemother. 1: Downloaded from on January 8, 2019 by guest