Quality Control Testing with the Disk Antibiotic Susceptibility Test of Bauer-Kirby-Sherris-Turck DONNA J. BLAZEVIC, M.P.H., MARILYN H. KOEPCKE, B.S., A JOHN M. MATSEN, M.D. Departments of Laboratory Medicine and Pediatrics, University of Minnesota Scliool of Medicine, Minneapolis, Minnesota 55455 ABSTRACT Blazevic, Donna J., Koepcke, Marilyn H., and Matsen, John M.: Quality control testing with the disk antibiotic susceptibility test of Bauer-Kirby- Sherris-Turck. Am. J. Clin. Pathol. 57: 592-5, 12. During two 6-month periods, daily quality control procedures for the single high potency disk diffusion antibiotic sensitivity test were monitored for 16 frequently used antibiotics using control strains of Escherichia coli and Staphylococcus aureus. During the first 6-month period disks were stored at 4 to C. Disks were stored in a frozen condition ( 20 C.) during the second 6-month period. Testing during the first 6-month period showed greater day-to-day variability than during the second 6-month period. Zone sizes were within a 4 mm. range in % of tests with E. coli for all antibiotics tested with the exception of sulfisoxazole, and were within a 4 mm. range for S. aureus in % of tests with the exception of lincomycin. The acceptable range for lincomycin has been modified to 6 mm. Testing was performed by 13 individuals, and the results demonstrate the practicality, accuracy, and reproducibility of this method when carried out by a well standardized method in the clinical laboratory. CONSIDERABLE EFFORT has been expended to develop a standardized practical means for determining the antibiotic susceptibility of microorganisms. The well-standardized and relatively simple method of Bauer, Kirby, Sherris, and Turck x has been widely accepted and is now being used in many clinical microbiology laboratories. Studies in our laboratory 2 have demonstrated the reproducibility of the test results when compared with results from the laboratory of Dr. John C. Sherris. Received April 13, 11; received revised manuscript June 11, 11; accepted for publication July 14, 11. 592 As in any biologic testing system, it is essential to monitor regularly all steps in this method in order to maintain the highest accuracy. This report outlines the methods used in our laboratory to carry out regular quality control monitoring of sensitivity results, and gives the results of dayto-day testing using control organisms. Materials and Methods Disk sensitivity testing was performed according to the method of Bauer and assoicates. 1 Commercially prepared 150 mm. plates containing Mueller-Hinton agar at a depth of 4 to 5 mm. were obtained from
May 12 QUALITY CONTROL OF DISK ANTIBIOTIC TESTING 593 23.. 22 AMPICILLIN LOT //i 00T3 l Q&slULJ^U- YEARi /ITI MAAH'S E. CQLI c H 21 C^-r^^rr 444 444 44-44-^ 44*44^- 44* 44* 44* 44--44 414444 444 44444*- 44-44*- 44* 44-44- 44 444-44 444 44-44r a. 20 "3 19-7-7 7-7-7 7-7-7 7-7-7-7-7-7 7-7- -7-7- -7*7--7-7- -7-7- -7-7- -7-7- -7-7- 7-7-7-7-7-7 -7-7-7 7-7-7 7-7-7 7-7- -7-7- -7-7- -7^- -7-7- - ^ - ^ ^ ^ ^^ 4^7-44 i-4*i 1 2 3 4 5 6 7 8 9 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 31 Day 25. 24 CEPHALOTHIN L0T#iO070 6 23. "vt 7 T t T t T ^^V *V~ ""^ y^* * T *^* -^ ~ T ~ T ~ -^ -7-7-7-7 77*r TTTT 777-7*T- 77-77" 77< 7-7 -7-7-77 777 777-777 7*77 "77-77 77 TT-T* 77 >77 77 7-7 77 77-77" -77 "ir 77 77-7 ^77, 77 T7? 77 ^"T" 77^ "T^H >77- "TTT "H- "H" H"7 7*77 7^ 7^ 7"^-T-H-7-7 7-77" J*1J I44 ^^ 1±- ^l~ ±1- ^J '-Z i-j 7-777 77 777 7-77 H - -^-^-^-^ - 1-1 - -^~ -^-^ -~^- 77 '~^ -^-^ni -44- -4*1 4*7 -r4-l 19 :. 2 3 4 5 6 7 8 9 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Z6 27 28 29 31 Day 26 25 24 23 2? 21 CHLORAMPHENICOL LOT #1 0C7O : : : : : : : : : :.'.''... i ; : : : :>:: : : : : ~i : : : :> : : : : ' ' '. '. '. * 1 2 3 4 5 6 7 8 9 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 31 Day FIG. 1. Quality control charts for disk susceptibility testing. Shaded areas indicate acceptable zone sizes. Daily results of zone sizes are indicated by solid black circles; results obtained with disks from new cartridges indicated by "x." '.V '.". : : : : : : : : : :< :: : : : : Baltimore Biological Laboratories (BBL). Antibiotic disks were also obtained from BBL. During the first 6-month test period all disks were stored at a temperature of 4 to C. in vials in screw-capped containers containing a dessicant. During the second 6-month test period, disks were stored in similar containers, but were held at a temperature of approximately 20 C. When disks were to be removed from the freezer, the dessicant jar was opened briefly to remove one vial of disks. This vial was then placed in a second jar containing dessicant, the jar was closed, and the vial was allowed to equilibrate to room temperature before the containers were opened. For daily use, disks were stored in a BBL disk dispenser with dessicant at 4 to C. Organisms used for quality control were single strains of Escherichia coli and Staphylococcus aureus. These organisms were originally isolated from clinical specimens and have been used in our laboratory for a number of years as controls for tube and agar dilution tests and for disk diffusion sensitivities. Their sensitivity to all antibiotics has remained stable. These organisms were kept at room temperature on trypticase soy agar (BBL) slants. New subcultures were used as needed, and frozen reference cultures were available if discrepant results occurred. Zones of inhibition were measured using
5 BLAZEVIC ET AL. A.J.C.I'. Vol..57 Table 1. Acceptable s of Inhibition Zones for 16 Antibiotics Using University of Minnesota Control Organisms Disk Antibiotic Content* E. coli\ S. aureus^ Chloramphenicol Colistin Erythromycin Gentamicin Kanamycin Lincomycin Methicillin Naladixic acid Nitrofurantoin Penicillin Polymyxin B Streptomycin Sulfisoxazole * Given in micrograms, except for penicillin and polymyxin B, which are given in units. t Diameter of zone of inhibition in millimeters, j = not done. a Fisher-Lilly zone reader and were recorded on control sheets each morning. Zone sizes for each antibiotic using the control organisms were tabulated, and the median zone size was taken as the value for establishing the quality control reference zone size. The quality control data charts are simi- Table 2. Results of Daily Quality Control Testing with Escherichia coli Antibiotic Chloramphenicol Colistin Gentamicin Kanamycin Naladixic acid Nitrofurantoin Polymyxin B Sulfisoxazole of No. Accepti- of No. % bility Times Within Within (mm.) Tested ±3 IS 2 5 0 0 0 192 237 140 165 170 1 190 2 18-22f 22-26 12-16 t 19-23 22-26 12-16 22-28 187 187 232 138 164 168 1 189 226 34-38 34-38 J 26-13-19 17-23 36-40 17-21 25-29 lar to the ones advocated by Dr. John C. Sherris. 8 A sample is shown in Figure 1. Zone sizes for each disk for the control organism used are marked on the charts each day by the technologist responsible for sensitivity testing on that particular day. These values are shown by the small dots on the chart. Thirteen technologists were involved in the tests on an alternating basis. When new disk cartridges were tested before being used for routine testing, the zone sizes obtained were recorded with an "X." Results Table 1 outlines the range of zone sizes deemed "acceptable" for our control organisms, as this range essentially contains all values falling within 2 standard deviations of the median. With all antibiotics tested except lincomycin and sufisoxazole, this range is mm. Table 2 outlines the results of the second 6 months of testing using these rather narrow limits of "acceptability" for the E. coli control organism. All tests using the aminoglycoside antibiotics gentamicin and kanamycin were within the established range, as were all tests with polymyxin B. Testing with ampicillin, naladixic acid, and nitrofurantoin was % within this mm. limit, and with the other antibiotics tested on a daily basis, testing was within the range at least % of the time. The one exception was testing with sulfisoxazole. By increasing the acceptable range to ±3 mm., fewer than 1% of tests fell outside those values. Table 3 contains the results of testing with the S. aureus control strain. Penicillin and streptomycin zone readings were % within the mm. range, ampicillin %, methicillin %, cephalothin 95%, and erythromycin and tetracycline %. Using a mm. range would have meant lincomycin would have been within those parameters in only 87% of tests. Using a
May 12 QUALITY CONTROL OF DISK ANTIBIOTIC TESTING 595 ±3 mm. range for lincomycin improves the acceptability to %. Results for the first 6-month period showed greater variability of values when disks were not stored frozen. Other undetermined factors may also have played some role in increasing variability. Table 4 gives values for percentages of tests falling within the above-named ranges for the two 6-month comparison periods. Values of ±3 mm. during the first period generally are lower than the values of mm. during the second period, especially in testing with Staphylococcus aureus using cephalothin, erythromycin, and penicillin. Lincomycin comparison figures are not available. Discussion There are many variables in the disk sensitivity method outlined by Bauer and associates, 1 as described above. The depth of the agar should be uniform, and can be routinely checked and corrected, if necessary. We have found that occasional lots Table 3. Results of Daily Quality Control Testing with Staphylococcus aureus Antibiotic Erythromycin Lincomycin Methicillin Penicillin Streptomycin of Acceptability (mm.) ±3 No. of Times Tested 164 172 171 167 168 No. Within 163 154 169 165 174 157 % Within 95 of plates, our own or commercially prepared, will have altered ph readings; therefore, we routinely check the ph of plates used. The potency of many antibiotics can be affected by changes in ph. The inoculum size is checked for each determination by using a turbidity standard which can be evaluated by its optical density in a simple spectrophotometer, as well as by visual perception of the growth on the plates. Table 4. Comparison of the two 6-month Study Periods: Percentages of Tests within Specified s with E. colt and Staphylococcus aureus Test Organisms First 6-month Period Second 6-month Period Organism Antibiotic No. of Times Tested %in ± 2 mm. %in ± 3 mm. No. of Times Tested %in ± 2 mm. Escherichia colt Chloramphenicol Colistin Gentamicin Kanamycin Naladixic acid Nitrofurantoin 256 315 6 178 255 226 391 88 79 84 91 85 87 70 96 95 88 192 237 140 165 170 2 Staphylococcus aureus Erythromycin Methicillin Penicillin Streptomycin 146 250 238 251 249 147 137 87 75 68 80 72 91 96 89 87 86 164 171 167 168 95
596 BLAZEVIC ET AL. A.J.C.P. Vol. 57 The reproducibility and accuracy of using a turbidity standard is improved by using a white background with black contrast lines held near a good light source. 4 Other factors which can be easily controlled are: the age of the uninoculated plates, the technic of inoculation, the temperature and time of incubation, the proper placement of disks, and the immediate incubation after placement of the disks. A realization that this method is standardized for testing with rapidly growing organisms (growth within 24 hr.) will help insure reliable interpretation. The aspect of the disk sensitivity test most difficult to control is the activity of the antibiotic disks themselves. Due to the necessity that stable organisms of known sensitivity (organisms should be sensitive to all antibiotics tested) be used for control of the disk method, we have used strains which have been shown over a period of several years to be very stable. It is not the purpose of this paper to advocate any use of our strains by others, but rather to stress the necessity for stable control organisms, and to point out the impressive reproducibility of the method when all the steps are accurately followed and the possible variables are well controlled. The advantage of having stable, widely available control strains will, hopefully, be obvious to the reader. The control strains used by the University of Washington have recently become available through the American Type Culture Collection (E. coli #25922 and 5. aureus #25923), and have been shown to be appropriate control strains. However, no matter what the source of control strains, it is necessary to establish an acceptable range of zone sizes for each antibiotic in order to be aware if results are at variance from the norm for the laboratory. Although it is not possible to perform a quality control check of the disk actually used on a given organism, the procedures used in this study indicate that one can maintain a most satisfactory means of evaluating the daily accuracy of the overall disk testing method. Having charts which clearly delineate the acceptable zone size ranges for each antibiotic allows the technologist performing the test to evaluate when technics have been accurate and when the antibiotic disks are appropriately active. The Sherris charts have been constructed to require a minimum of writing, and any member of the laboratory staff can quickly peruse daily test results. If all disks on a given plate give zone sizes out of a range, it is likely that a problem has arisen with the medium or the technic of the test. In this case all tests done that day should be repeated. In our experience this type of a problem has never occurred, although we have, by using control methods for other variables, discovered problems which might have led to errors with all disks being tested. If one disk on a plate gives a zone which is out of range, it may be that this represents a single inactive disk in the cartridge or vial. However, the possibility that there are other disks with altered activity in that container is high enough to warrant discarding the container. If one is to use a new cartridge, ideally a disk from that cartridge should be checked for activity prior to use for routine testing in the laboratory. In this study, with the exception of sulfisoxazole, there were very few readings outside the mm. range with E. coli. It should be stressed that in doing the disk sensitivity test with sulfonamides, Mueller- Hinton agar is an acceptable medium on which to test the sulfonamides, but that organisms being tested will grow through three or four generations before being inhibited. Therefore, the recommendations of Bauer and associates 1 are that one read 80% inhibition to determine zone size. It is logical, then, that these zone measure-
May 12 QUALITY CONTROL OF DISK ANTIBIOTIC TESTING 5 ments would be less reproducible than those with antibiotics which give complete inhibition and sharply defined zones of inhibition, especially if a number of different individuals are interpreting zone sizes. To ascertain the variability which would occur if only one individual were doing the reading of zone sizes, one of us (M. H. K.) read sulfisoxazole zones on 52 different quality control checks with E. coli. The zone sizes were within mm. 49 times, for 96% agreement. Were this criterion used for the readings by all technologists, the result would have been 165 acceptable readings out of 190 tests, for an acceptability of only 87%. Therefore, it was felt that in this instance the test itself was not at variance, but merely the readings, and that an acceptability range of ±3 mm. was appropriate for maintaining accurate control. There were a few more zone readings outside the mm. range for S. aureus for most antibiotics when compared with E. coli figures. This probably occurs because with this organism zones are not as clearly outlined as they are with E. coli. The acceptable zone size range for lincomycin was broadened due to the greater variability of test results with this organism. A check done again by having one of us (M. H. K.) read a large number of tests showed 45 of 49 tests (92%) to be within a mm. range. It was felt that in order to obviate unnecessary retesting the acceptable zone sizes for lincomycin in our system should be ±3 mm. We feel that the overall results of this study are reassuring, and demonstrate the impressive reproducibility of the Bauer- Kirby-Sherris-Turck method when performed in a well standardized manner. In a large laboratory such as ours, daily control testing maintains an effective check on problems which might arise, takes relatively little extra effort, and encourages a high standard of performance on the part of the technologists responsible for that day's testing. The results of the comparison of the two 6-month periods indicate that reproducibility and accuracy were improved when the disks were stored at 20 C. before use. This is especially important with the penicillins and cephalosporins, but by maintaining standard conditions for all disks, the overall reproducibility and control of the test are enhanced. Acknowledgment. The staff of the Diagnostic Microbiology Laboratory cooperated in this study. References 1. Bauer AW, Kirby WMM, Sherris JC, et al: Antibiotic susceptibility testing by a standardized single disk method. Am J Clin Pathol 45: 493-496, 1966 2. Matsen JM, Koepcke MJH, Quie PG: Evaluation of the Bauer-Kirby-Sherris-Turck singledisc diffusion method of antibiotic susceptibility testing. Antimicrob Agents Chemother 1969-10, pp 445-453 3. Sherris JC: Personal communication, 1968 4. Stemper JE, Matsen JM: Device for tubidity standardizing of cultures for antibiotic sensitivity testing. Appl Microbiol 19:15-16, 10