ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Mar. 1982, p. 412-415 Vol. 21, No. 3 0066-4804/82/030412-04$02.00/0 Examination of Pseudomonas aeruginosa-Gentamicin Discrepancies Encountered in an Autobac I-Disk Diffusion Comparison JOAN B. MAYO, TIMOTHY E. KIEHN,* BRIAN WONG, EDWARD M. BERNARD, AND DONALD ARMSTRONG Diagnostic Microbiology Laboratory, Memorial Hospital, and Infectious Disease Service, Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, New York 10021 Received 31 August 1981/Accepted 29 December 1981 Seven Pseudomonas aeruginosa strains found to be susceptible to gentamicin by the Autobac I system and resistant by the Bauer-Kirby disk diffusion method were tested for the presence of mixed populations of cells. Double zones of inhibition randomly appeared on gentamicin disk diffusion plates, and susceptible and resistant subpopulations were isolated from these plates. Growth kinetic Downloaded from aac.asm.org by on June 24, 2010 studies of separated strains and mixed subpopulations indicated that the suscepti- ble organisms grew rapidly and were read as susceptible at 5 h with the Autobac I system. Resistant organisms entered a sustained lag phase and did not achieve sufficient turbidity to be read as resistant with the Autobac I system before 6 h. Thus, a false reading of susceptible could be obtained with the Autobac I system, depending on the ratio of susceptible organisms to resistant organisms selected for testing. A mixed susceptible and resistant population could be recognized by extending the incubation time of the Autobac I cuvette or by using other susceptibility methods to test isolates with light-scattering indexes of <1.0. The Autobac I system (General Diagnostics) difficulty identifying methicillin-resistant staph- is a semiautomated system for rapid (3 to 5 h), ylococci and gentamicin-resistant P. aeruginosa semiquantitative susceptibility testing of bacte- strains (7). These organisms were rarely seen in ria. In an initial collaborative study, the per- this country at that time. Subsequent reports formance of this system was shown to have a have shown that the Autobac I system indeed high overall level of agreement (91.5%) with the had difficulty detecting methicillin-resistant disk diffusion method (7). However, a number of Staphylococcus aureus (1, 3). In one study (1), discrepancies have been reported for several the authors were able to demonstrate that there organism-antibiotic combinations (1, 3, 4, 6, 7). were mixed populations of susceptible and resis- In a preliminary study in our laboratory (T. E. tant organisms within their S. aureus cultures. Kiehn, J. B. Mayo, and D. Armstrong, Abstr. When we began to study the Pseudomonas Annu. Meet. Am. Soc. Microbiol. 1979, C75, p. isolates that gave discrepant susceptibility test 322), the antibiotic susceptibilities of 1,033 results, the following two observations suggest- gram-negative clinical isolates were determined ed that we might also be working with mixed by both Autobac I and Bauer-Kirby disk diffu- populations: (i) double zones of growth were sion methods. We encountered 42 discrepancies found around some of the gentamicin disks on which were confirmed by repeated testing. The the disk diffusion plates, and (ii) the results of most prevalent discrepancy was observed when the tests often changed back and forth between we tested the susceptibility of Pseudomonas susceptible and resistant on repetitive testing. In aeruginosa isolates to gentamicin; the Autobac I this paper we report the results of studies under- system indicated that the organisms were sus- taken to explain the discrepancies which we ceptible, whereas the disk diffusion and broth observed when we performed tests on the sus- dilution methods indicated that the organisms ceptibility of P. aeruginosa to gentamicin. were resistant. These inconsistencies when the Autobac I system was used were particularly MATERIALS AND METHODS disturbing because gentamicin is used common- Organisms. Suspensions of the seven P. aeruginosa- ly to treat Pseudomonas infections in our hospi- gentamicin discrepant isolates were preserved for fur- tal. ther study in sterile, defibrinated sheep blood at -70°C The authors of the original collaborative study until further testing. The organisms were thawed and suggested that the Autobac I system might have subcultured on Columbia agar plates containing 5% 412 VOL. 21, 1982 AUTOBAC I-DISK DIFFUSION DISCREPANCIES 413 sheep blood. The plates were incubated overnight at isms were determined by comparing the number of 35°C. The control strains used throughout the study colonies on plates containing gentamicin with the were Escherichia coli ATCC 25922 and ATCC 29194 number of colonies on the antibiotic-free plates. and P. aeruginosa ATCC 27853. Susceptibility test methods. The Autobac I system RESULTS was used according to the instructions of the manufac- turer, except where indicated. The susceptibility pro- A series of gentamicin disk diffusion plates cedures outlined in the original collaborative evalua- were inoculated with each discrepant Pseudo- tion of the Autobac I system were followed (7). The monas isolate in an attempt to demonstrate Autobac I photometer automatically measured the mixed susceptible and resistant populations. extent of inhibition of growth by the antimicrobial Each plate exhibited a susceptible zone of inhi- agent compared with the growth control. This ratio was expressed as a numerical value between 0 and 1, bition, no zone of inhibition, or double zones of called the light-scattering index (LSI). An LSI of .0.6 inhibition. The mixed nature of the isolates was used to define susceptibility, and an LSI of <0.5 showing double zones of inhibition was con- was used to define resistance. Values ranging from firmed by Autobac I, disk diffusion, and broth 0.51 to 0.59 were defined as showing intermediate dilution susceptibility test methods. The genta- susceptibility. micin-susceptible subpopulations gave LSI The standardized agar disk diffusion method was readings of 0.68 to 1.0, disk diffusion zone sizes performed as recommended by the National Commit- of 15 to 17 mm, and minimal inhibitory concen- tee for Clinical Laboratory Standards (5). trations of 2 or 4 ,ug/ml. The gentamicin-resis- Minimal inhibitory concentrations were determined Downloaded from aac.asm.org by on June 24, 2010 by using a microtiter broth dilution method, which was tant subpopulations gave LSI readings of 0.19 to performed according to international collaborative 0.49, disk diffusion zone sizes of zero, and study standards (2). Reference powders for each anti- minimal inhibitory concentrations of 32 or 64 ,ug/ microbial agent tested were obtained from the drug ml. manufacturers. These findings suggested that the discrepan- Separation of susceptible and resistant Pseudomonas cies were due to random selection of the suscep- subpopulations. Susceptible and resistant organisms tible or resistant organisms in a mixed popula- from populations of the seven Pseudomonas isolates tion. The short incubation period with the studied were separated by using the disk diffusion Autobac I system might not allow the resistant method. Standardized inocula containing 108 colony- forming units per ml were streaked onto Mueller- subpopulation in a mixed inoculum to grow to Hinton plates on which disks containing 10 ,ug of sufficient density for proper interpretation. To gentamicin were placed. These plates were incubated test this, the growth kinetics for the separated at 35°C overnight. If double zones of inhibition were subpopulations of Pseudomonas in the presence observed, colonies within 10 mm of the disk and of gentamicin elution disks were determined colonies more than 15 mm from the disk were selected during an extended incubation period by using and tested individually by the Autobac I, disk diffu- the Autobac I system. The results for a repre- sion, and broth dilution methods. sentative isolate (isolate 33) are shown in Fig. 1. Growth kinetics of susceptible and resistant subpopu- At 5 h, the susceptible strain had reached a lations. The growth rates of the separated susceptible and resistant strains of P. aeruginosa were determined density sufficient for reading on the Autobac I by using the Autobac I system. The organisms were system. The maximum LSI of 1.0 and a corre- incubated in Autobac I cuvettes in the presence of sponding interpretation of susceptible persisted gentamicin elution disks for up to 10 h, and LSI during the entire incubation period. The resis- readings were recorded at zero time and at hourly tant strain did not grow to sufficient density for a intervals. reading until after 6 h of incubation. LSI read- Growth kinetic studies were then conducted with ings decreased during continued incubation and mixtures of the susceptible and resistant subpopula- gave interpretive readings of increased resist- tions of the P. aeruginosa isolates. Each mixture was ance during the incubation period. These results prepared by combining 2 ml of each of the subpopula- tions, which had been standardized individually with demonstrated that the resistant strain grew more the Autobac I system. Subsequent dilutions were slowly than the susceptible strain and would not made in Eugonic broth to yield final inoculum concen- have been detected after 5 h of incubation. trations of 106 to 2 x 106 colony-forming units per ml. The growth rate of an experimental mixture of The effect of extended Autobac I incubation on the susceptible and resistant strains was determined susceptibility results was determined as described as described above for each Pseudomonas iso- above. In addition, the relative numbers of susceptible late. The Autobac I results for isolate 33 are and resistant organisms in the mixture at zero time and shown in Fig. 1. The mixture of susceptible and at each incubation interval were determined. Samples (100 ,uJ) were taken from the cuvettes, and serial 10- resistant subpopulations could be read at 5 h, fold dilutions were made in Mueller-Hinton broth; 0.1- similar to the separated susceptible strain, and ml samples were placed onto Mueller-Hinton plates LSI values decreased during the extended incu- with and without 10 ,g of gentamicin per ml. After 24 bation period, similar to the separated resistant h the number of colonies on each plate was counted, strain. The mixture was read by the Autobac I and the proportions of susceptible and resistant organ- system as susceptible at 5 h and resistant at 7.5 414 MAYO ET AL. ANTIMICROB. AGENTS CHEMOTHER. a) .8 S ._ 7 4- c)4 c3- R\ .2- 1 2 3 4 5 6 7 8 0 8 Hours Hours FIG. 1. Autobac I growth kinetics for susceptible FIG. 2. Number of susceptible (0) and resistant strains (O), resistant strains (A), and a mixture of (A) organisms present during growth of a mixture of the two strains of Pseudomonas isolate 33 when susceptible and resistant strains (O) of Pseudomonas Downloaded from aac.asm.org by on June 24, 2010 isolate 33 in the presence of gentamicin elution disks. gentamicin was tested with the Autobac I method. S, I, and R indicate the susceptible, intermediate, and CFU, Colony-forming units. resistant LSI breakpoints, respectively, for Autobac I susceptibility determinations. nas isolates indicated that mixed susceptible and resistant populations might be a factor. Mixed h. The resistant strain was not detected at the 5- populations have been reported in Autobac I h reading because of its slower growth. studies with cultures of methicillin-resistant S. More detailed growth curves were determined aureus (1). by removing samples from each cuvette and. Our results suggest that with Pseudomonas counting the susceptible and resistant organisms and gentamicin and possibly with other combi- at hourly intervals. Figure 2 shows the results nations, the prolonged lag phase of the resistant for isolate 33. The resistant population de- strains results in a failure of these strains to creased during the first hour, entered a sustained manifest themselves in a mixed culture during lag period, and attained an exponential growth the short incubation period required for Autobac rate after 3 h. In contrast, the susceptible popu- I susceptibility testing. Thus, a "false" reading lation exhibited a brief lag period and rapidly of susceptible could be obtained with the Auto- entered the exponential growth phase. After 5 h, bac I system, depending on the ratio of suscepti- there was an approximate 2-log-greater number ble strains to resistant strains selected for test- of susceptible organisms present. ing. An extended period of incubation of the Subpopulations of susceptible and resistant Autobac I cuvette has been suggested as a organisms were separated from within popula- remedy for the erroneous results obtained with tions of the remaining six Pseudomonas-genta- S. aureus and the penicillins (3). Our results also micin discrepant isolates. Mixtures of the sepa- indicate that a decrease in the LSI upon extend- rated susceptible and resistant subpopulations ed incubation of the cuvette could serve as a for each isolate gave double zones of inhibition marker for detecting such a mixture. However, and decreased LSI values when the Autobac I this requirement would compromise one of the incubation period was extended. principle advantages of the Autobac I system, the early determination and reporting of suscep- DISCUSSION tibility results. Several factors have been proposed to explain Another indicator of a mixed resistant and discrepancies between the results obtained with susceptible population could be an initial LSI of the Autobac I and agar diffusion methods. These <1.0. Funnell and Guinness (3) showed that include differences in types of basal media, their staphylococci that were methicillin suscep- length of incubation, the narrow intermediate tible, as determined by disk diffusion, invariably range and greater sensitivity of the Autobac I had LSIs of >0.9, whereas their staphylococci system, and the technical and biological variabil- that were methicillin resistant had LSIs of ities inherent in comparisons between any two <0.70. Six of our seven discrepant pseudomo- test systems (7). In this study, the presence of nads had initial LSIs between 0.6 and 0.97, double zones of inhibition on agar diffusion whereas 152 (96%) of our truly gentamicin- plates with gentamicin-susceptible Pseudomo- susceptible Pseudomonas sp. isolates exhibited VOL. 21, 1982 AUTOBAC I-DISK DIFFUSION DISCREPANCIES 415 LSI readings of 1.0. Pseudomonas isolates ex- allow an understanding of the nature of some of hibiting LSIs of 0.6 to 0.9 for gentamicin might the limitations of the Autobac I method. contain mixed susceptible and resistant strains and should be retested by conventional disk ACKNOWLEDGMENTS diffusion, minimal inhibitory concentration The assistance of the technologists of the Diagnostic Micro- methods, or by extended incubation of the Auto- biology Laboratory, Memorial Hospital, is gratefully acknowl- bac I cuvette. edged. Helen Rigopoulos was very helpful in the preparation of the manuscript. In our original evaluation of the Autobac I system, 42 significant discrepancies were ob- LITERATURE CITED served during testing of 1,033 gram-negative 1. Cleary, T. G., and D. Maurer. 1978. Methicillin-resistant isolates. The discrepancies which were of great- Staphylococcus aureus susceptibility testing by an auto- est concern from a clinical perspective, involv- mated system, Autobac I. Antimicrob. Agents Chemother. ing Pseudomonas and gentamicin, were studied 13:837-841. in detail. We found that these isolates contained 2. Erlccson, H. M., and J. C. Sherris. 1971. Antibiotic sensi- tivity testing: report of an international collaborative study. subpopulations with different susceptibilities. Acta Pathol. Microbiol. Scand. Sect. B. 217(Suppl.):1-90. The other 35 discrepant organism-antibiotic 3. Funnell, G. R., and M. D. G. Guinness. 1979. Australian combinations included Pseudomonas with car- evaluation of Autobac I with suggested interpretive and beniciilin and chloramphenicol, Escherichia and technical modifications. Antimicrob. Agents Chemother. 16:255-261. Enterobacter with ampicillin, cephalothin, and 4. Harris, P. C., and L. B. Sealey. 1980. Two-hospital study chloramphenicol, Serratia with carbenicillin and Downloaded from aac.asm.org by on June 24, 2010 of Staphylococcus aureus susceptibility to penicillin and chloramphenicol, Klebsiella with carbenicillin, ampicillin by Autobac I. Antimicrob. Agents Chemother. cephalothin, and gentamicin, and Proteus and 18:922-925. 5. National Committee for Clinical Laboratory Standards. Citrobacter with cephalothin. In preliminary 1979. Performance standards for antimicrobial disc suscep- studies, we found that 32 of these organisms also tibility tests. Approved standard: ASM 2. National Com- exhibited double zones of inhibition or de- mittee for Clinical Laboratory Standards, Villanova, Pa. creased LSIs after extended incubation when 6. Stubbs, K. G., and K. Wicher. 1977. Laboratory evaluation of an automated antimicrobial susceptibility system. Am. J. the particular antibiotic was tested. Clin. Pathol. 68:769-777. The Autobac I system interprets a dynamic 7. Thornsberry, C., T. L. Gavan, J. C. Sherris, A. Balows, process by sampling a culture at a fixed interval J. M. Matsen, L. D. Smith, F. Schoenknecht, L. D. Thrupp, early in the growth curve. The results obtained and J. A. Washington II. 1975. Laboratory evaluation of a rapid automated susceptibility testing system: report of a depend on the nature of the population and its collaborative study. Antimicrob. Agents Chemother. growth kinetics. Attention to these factors might 7:466-480.
Pages to are hidden for
"Examination of Pseudomonas aeruginosa-Gentamicin"Please download to view full document