Diagnostic Microbiology and Infectious Disease 38 (2000) 43–50
www.elsevier.com/locate/diagmicrobio
Evaluation of antibiotic synergy against Acinetobacter baumannii: a comparison with Etest, time-kill, and checkerboard methods Charles R. Bonapace, Roger L. White, Lawrence V. Friedrich, John A. Bosso,* Anti-Infective Research Laboratory, College of Pharmacy, Medical University of South Carolina, Charleston, South Carolina, USA Received 20 January 2000; received in revised form 15 June 2000; accepted 15 June 2000
Abstract Acinetobacter baumannii is becoming increasingly resistant to antibiotics, often requiring combination therapy. Numerous methods exist to detect the presence of in vitro synergy with the time-kill and checkerboard tests being widely used. The Epsilometer test (E test) is a new method that is less labor intensive, but has not been evaluated using a wide range of antimicrobials and organisms. We assessed synergy using the time-kill and checkerboard tests and compared the results to the E test method using 10 clinical isolates of A. baumannii. Antimicrobial combinations evaluated consisted of trovafloxacin or tobramycin in combination with cefepime or piperacillin. Synergy was detected with all combinations by either the checkerboard or time-kill method. Synergy was not detected by the Etest method. The agreement between the time-kill test and Etest method was 72% (range 42–97%); for the time-kill and checkerboard tests, agreement was 51% (range 30 – 67%). The Etest method appears promising although further testing should be performed with additional antimicrobial agents and organisms. © 2000 Elsevier Science Inc. All rights reserved.
1. Introduction Acinetobacter spp. have become important pathogens, particularly in intensive care units and are associated with a variety of nosocomial infections including pneumonia, bacteremia, urinary tract infections, endocarditis, skin and wound infections, and meningitis (Bergogne-Be´re´zin, 1987, 1991, 1996, Ng, 1989). Numerous genomic species of Acinetobacter have been identified. However, it is recognized that Acinetobacter baumannii is the species most responsible for outbreaks of nosocomial infections (Bouvet, 1987). Antibiotic resistance with Acinetobacter baumannii is now common and has increased dramatically over the last two decades (Bergogne-Be´re´zin, 1996). Most strains are resistant to older antibiotics including extended-spectrum penicillins, 1st and 2nd generation cephalosporins, amino-
* Corresponding author. Tel.: ⫹1-843-792-8501; fax: ⫹1-843-7921617. E-mail address:
[email protected] (J.A.Bosso). Presented in part at the American Society for Microbiology General Meeting, Atlanta, Georgia May 1998, Abstract #A-20
glycosides, and tetracyclines (Bergogne-Be´re´zin, 1996). While some strains remain susceptible to 3rd generation cephalosporins, resistance to drugs such as tobramycin, amikacin, fluoroquinolones, and imipenem has been encountered (Tankovic, 1994). Due to high rates of resistance associated with Acinetobacter spp., combination therapy is often employed for the treatment of infections. Despite resistance to these older agents, commonly used regimens often include an extended-spectrum penicillin, 3rd generation cephalosporin, or imipenem in combination with either an aminoglycoside or a fluoroquinolone (Bergogne-Be´re´zin, 1996). Synergy is a potential benefit of combination antibiotic therapy for Acinetobacter infections. However, synergy is not an all-or-none phenomenon and it may need to be assessed on an isolate-by-isolate basis. While numerous methods used to detect in vitro synergy between antibiotics have been described, the checkerboard and time-kill tests are the most widely used techniques. The checkerboard test, a gauge of inhibitory activity, is a relatively easy test to perform. The time-kill test assesses bactericidal activity but is time consuming and labor-intensive. Thus, neither test is routinely used in the clinical setting. A relatively new agar diffusion method for antimicrobial susceptibility testing is the Epsilometer (Etest). We have previously described a method of synergy testing comparing
0732-8893/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 2 - 8 8 9 3 ( 0 0 ) 0 0 1 6 3 - 2
44
J.A.Bosso et al. / Diagnostic Microbiology and Infectious Disease 38 (2000) 43–50
an Etest method to the time-kill and checkerboard tests with E. coli, E. cloacae, P. aeruginosa, and S. aureus using either tobramycin or ciprofloxacin plus cefepime or ceftazidime (White, 1996). Despite the fact that each of these methods uses different endpoints, agreement of qualitative interpretation was demonstrated among the methods. The purpose of the current study was to assess synergy using trovafloxacin or tobramycin with cefepime or piperacillin and to compare the agreement between the checkerboard test, time-kill test, and Etest method using 10 clinical isolates of Acinetobacter baumannii.
2. Materials and methods Antimicrobial agents. Laboratory grade standard powders of cefepime hydrochloride (lot # 6R142, Bristol-Myers Squibb, Syracuse, NY), piperacillin sodium (lot # 85H0838, Sigma-Aldrich, St. Louis, MO), and trovafloxacin mesylate (lot # 32438 –280-1M, Pfizer Inc., Groton, CT) were obtained. Laboratory grade tobramycin solution 10,000 g/mL (lot # X00899 SI08183, Eli Lilly and Co., Indianapolis, IN) was used. All of the above standards were used for the broth MIC determination, checkerboard test, and timekill test. Stock solutions of each antimicrobial agent were prepared using deionized water and diluted in cation-adjusted Mueller-Hinton broth. Etest strips of cefepime (lot #s B70308 and B70282), piperacillin (lot #s B63042 and B62710), tobramycin (lot #s B62937 and B62399), and trovafloxacin (lot #s B63147 and B51990) (AB Biodisk, Solna, Sweden) were used for Etest MIC determination and the Etest method of synergy testing. Microorganisms and media. Ten clinical isolates of Acinetobacter baumanni isolated from non-urine sources from 10 different patients were studied. Strains were isolated between 1993 and 1996. All strains were identified by the clinical microbiology laboratory using standard methods and stored at –70°C in Microbank® vials (lot # 31538, Pro-Lab Diagnostics, Austin, TX). Prior to use, each strain was transferred onto a Mueller-Hinton II agar plate, grown overnight, and inoculated into cation-adjusted Mueller-Hinton broth. Pseudomonas aeruginosa ATCC 27853 was included as a quality control strain for all MIC testing. Mueller-Hinton broth (lot # 84585JG, Difco Laboratories, Detroit, MI) supplemented with CaCl2 and MgCl2 to produce final concentrations of 25 g/mL and 12.5 g/mL, respectively, was prepared according to the National Committee for Clinical Laboratory Standards (NCCLS) guidelines (National Committee for Clinical Laboratory Standards 1993). Fifteen cm Mueller-Hinton II agar plates (lot #s C1RWCW and K4RBSI, BBL, Cockeysville, MD) were used for the Etest MIC determination and the Etest method of synergy testing, whereas 10 cm Mueller-Hinton II agar plates (lot #s A2RCTC, B4RFRJ, and D2RGPT, BBL, Cockeysville, MD) were used for enumeration of the final inoculum and surviving colonies for the time-kill test.
MIC determination. The MICs of piperacillin, cefepime, tobramycin, and trovafloxacin were determined by broth microdilution and Etest. Broth microdilution customary (e.g., 1, 2, 4 g/mL) and intermediate (e.g., 1.5, 3, 6 g/ml) MICs were each performed in replicates of four according to NCCLS guidelines (NCCLS, 1993). The lowest modal MIC (customary or intermediate) was used to design the concentration range of the checkerboard and time-kill tests. Microtiter trays were prepared using an automated 96-well inoculator (Sensititre, Chelsea Instruments LTD, England) and two-fold dilutions performed using an automated diluter (Autodilutor III, Dynatech Laboratories Inc., Alexandria, VA). The concentration range tested was 0.03 to 1536 g/mL for cefepime, 0.09 to 4096 g/mL for piperacillin, 0.00009 to 4 g/mL for trovafloxacin, and 0.02 to 768 g/mL for tobramycin. Trays were sealed and frozen at –70°C and used within 24 h. Each isolate was allowed to grow in Mueller-Hinton broth at 35°C for three hours to achieve logarithmic growth. The inoculum was then prepared in 0.9% saline at 35°C by matching the optical density to a 0.5 McFarland standard and further diluting in cationadjusted Mueller-Hinton broth. The final inoculum was approximately 5 ⫻ 105 CFU/mL and was verified in duplicate using a spiral plater (Spiral Systems, Cincinnati, OH). The microtiter trays were sealed and incubated for 18 h at 35°C in ambient air. The MIC was defined as the lowest drug concentration that prevented turbidity as read using a magnified lighted microplate reader (Sensititre, Chelsea Instruments LTD, England). MICs were also determined in triplicate using Etest strips (AB Biodisk, Solna, Sweden) and the modal MICs used for the Etest synergy method. Each isolate was matched to a 0.5 McFarland standard and streaked on a 15 cm MuellerHinton II agar plate as directed by the manufacturer. Four Etest strips (one strip for each antimicrobial) were placed on each agar plate. The concentration range on the Etest strips was 0.016 –256 g/mL for piperacillin, cefepime, and tobramycin and 0.002–32 g/mL for trovafloxacin. Agar plates were incubated for 18 h at 35°C in ambient air. The MIC was defined as the concentration at which the ellipse intersected the concentration scale imprinted on the Etest strip and interpreted according to the manufacturer’s guidelines. Synergy testing. Synergy testing was performed using the checkerboard test, time-kill test, and Etest methods. The liquid media methods, the checkerboard and time-kill tests, were performed simultaneously for each isolate to prevent interday variation in the final inoculum and length of time the antimicrobial solutions were frozen. The following antimicrobial combinations were tested with all 3 methods: cefepime ⫹ tobramcyin, cefepime ⫹ trovafloxacin, piperacillin ⫹ tobramycin, and piperacillin ⫹ trovafloxacin. The fractional inhibitory concentration index (FIC index) was used to interpret the checkerboard test and Etest method and calculated as follows: FIC index ⫽ FIC of drug A ⫹ FIC of drug B
J.A.Bosso et al. / Diagnostic Microbiology and Infectious Disease 38 (2000) 43–50
The FIC of drug A and drug B was calculated as follows FIC of drug A ⫽ MIC of drug A in combination/MIC of drug A alone and FIC of drug B ⫽ MIC of drug B in combination/MIC of drug B alone Synergy was defined as an FIC index ⱕ 0.5, additivity/ indifference was defined as an FIC index ⬎ 0.5 to 4, and antagonism was defined as an FIC index ⬎ 4 (Eliopoulos, 1996). Checkerboard test. Checkerboard synergy testing was performed in triplicate in 96-well microtiter trays using an 8-by-8 well configuration. Wells not containing antibiotic were used as positive growth controls to assess the presence of turbidity. Negative growth controls were performed in a separate microtiter tray. MICs were determined prior to performing the checkerboard test and not repeated the same day. Dilutions of each antimicrobial were performed using an automated dilutor with concentrations ranging from 0.031⫻ MIC to 4⫻ MIC. Once prepared, microtiter trays were immediately frozen at –70°C and used within 96 h. The initial inoculum, prepared as previously described, was approximately 5 ⫻ 105 CFU/mL and was verified after plating in duplicate using a spiral plater. Microtiter trays were sealed and incubated for 18 h at 35°C in ambient air. The FIC index of each non-turbid well along the turbidity/non-turbidity interface was calculated using the corresponding concentration of drug A and drug B in combination for that well. The mean FIC index of all non-turbid wells along the turbidity/non-turbidity interface was then calculated. The average of the mean FIC indices from the triplicate microtiter trays were used to categorize results as synergy, additivity/indifference, and antagonism. Time-kill test. Time-kill tests were performed in glass tubes with 30 mL of cation-adjusted Mueller-Hinton broth. Each isolate was tested against antimicrobial agents alone at concentrations of 0.25⫻ MIC and 2⫻ MIC and in combination at these concentration to MIC ratios. As an example of a combination using cefepime and tobramycin, each drug was tested at 0.25⫻ MIC and 2⫻ MIC by itself in addition to the following: cefepime 0.25⫻ MIC ⫹ tobramcyin 0.25⫻ MIC cefepime 0.25⫻ MIC ⫹ tobramycin 2⫻ MIC cefepime 2⫻ MIC ⫹ tobramycin 0.25⫻ MIC cefepime 2⫻ MIC ⫹ tobramcyin 2⫻ MIC Thus, 16 different drug/MIC combinations were tested for each isolate. However, the MICs of trovafloxacin for 5 isolates (#1, 6, 7, 8, and 10) exceeded its solubility in cation-adjusted Mueller-Hinton broth. Thus, synergy testing in liquid media (time-kill and checkerboard tests) was not performed for these isolates using combinations that included trovafloxacin. Hence, only 8 different drug/MIC combinations were tested for these isolates (4 combinations each of cefepime or piperacillin and tobramycin). Antimicrobial solutions for the time-kill test were stored in sterile polypropylene containers at –70°C for less than
45
96 h prior to synergy testing. Following thawing at room temperature, antimicrobial solutions were transferred into 30 mL glass tubes and inoculated with each isolate. The final inoculum was approximately 5 ⫻ 105 CFU/mL and was verified after plating in duplicate using a spiral plater. Each tube was agitated using an orbital shaker (150 RPM) in a microbiologic warm room at 35°C for 24 h. A 500 L sample was removed from each tube in duplicate at 0 and 24 h for colony count enumeration. Serial tenfold dilutions of these samples were performed in 0.9% saline at room temperature. Fifty microliters of undiluted and diluted samples were plated onto 10 cm antibiotic-free Mueller-Hinton II agar plates using a spiral plater, incubated at 35°C in ambient air for 18 h, and surviving colonies counted. The mean colony count (CFU/mL) from duplicate samples was used in the assessment of synergy. Colony counts at 24 h that were less than the limit of quantitation with the spiral plater (4 ⫻ 102 CFU/mL) were recorded as 4 ⫻ 102 CFU/ mL. The effect of the antimicrobial combinations were interpreted as follows: synergy was defined as at least a 100-fold increase in killing at 24 h with the combination compared to the most active single agent and at least 100-fold killing of the initial inoculum. Antagonism was defined as at least a 100-fold increase in colony count at 24 h with the combination when compared to the most active drug alone. Additivity/indifference was defined as any other scenario not meeting the criteria for either synergy or antagonism. A formal study to detect the presence of antibiotic carryover was not performed. However, the potential for antibiotic carryover was assessed for each individual drug in addition to each drug combination by calculating a ratio of the corrected colony count from a sample diluted in 0.9% saline to the corrected colony count from a sample diluted one tenfold dilution less in 0.9% saline. Ratios with a value greater than 1 were potentially suggestive of the presence of antibiotic carryover. Plotting the ratio vs. each dilution plated revealed no evidence that antibiotic carryover may have been present. In only a few instances were undiluted samples used for the enumeration of surviving colonies. Etest. The Etest method was performed in duplicate. Fifteen cm Mueller-Hinton II agar plates were streaked with each isolate that had been matched to a 0.5 McFarland standard. As shown in Fig. 1, the second strip was placed on the agar plate at a 90° angle to the respective MIC of each antimicrobial agent as previously described (White, 1996). When the MIC exceeded the concentration range on one or both of the Etest strips, the strips were crossed at the highest concentration present on the respective Etest strip. The FIC index was then calculated using an MIC one twofold dilution above the highest concentration on the Etest strip. This was necessary for trovafloxacin in 5 isolates (#1, 6, 7, 8, and 10; see Table 1) and in 4 isolates for piperacillin (#6, 7, 8, and 10). The agar plates were incubated at 35°C in ambient air for 18 h and the MIC of each antimicrobial in combination was read as previously described. The effect of the
46
J.A.Bosso et al. / Diagnostic Microbiology and Infectious Disease 38 (2000) 43–50
Fig. 1. Placement and interpretation of Etest strips for the Etest method of synergy testing
antimicrobial combinations was interpreted using the mean FIC index from the duplicate set of agar plates. Assessment of agreement among methods. Results from each synergy testing method were compared based on the interpretive category (synergy, additivity/indifference, and antagonism). Agreement between methods was defined as both methods having the same interpretive category, major disagreement was defined as one method displaying synergy and the other method antagonism, and minor disagreement as all other possibilities. 3. Results The broth microdilution and Etest modal MICs of the clinical isolates for the antimicrobials are shown in Table 1. Table 1 Model broth microdilution and Etest MIC values (g/mL) Isolate Piperacillin
#1 #2 #3 #4 #5 #6 #7 #8 #9 #10
Cefepime
Tobramycin
Trovafloxacin
Broth Etest
Broth
Broth
Etest
Broth Etest
8.00 16.00 192.00 16.00 32.00 256.00 1536.00 3072.00 16.00 1536.00
16.00 16.00 32.00 16.00 48.00 ⬎256.00 ⬎256.00 ⬎256.00 24.00 ⬎256.00
16.00 1.00 16.00 2.00 8.00 96.00 16.00 24.00 3.00 24.00
Etest
24.00 32.00 48.00 ⬎4.00 ⬎32.00 2.00 1.00 1.00 0.05 0.09 4.00 1.50 1.00 0.06 0.13 3.00 0.75 1.00 0.05 0.09 6.00 1.00 1.00 0.06 0.25 48.00 6.00 3.00 ⬎4.00 ⬎32.00 24.00 12.00 8.00 ⬎4.00 ⬎32.00 24.00 16.00 8.00 ⬎4.00 ⬎32.00 3.00 0.25 0.50 0.03 0.06 32.00 6.00 4.00 ⬎4.00 ⬎32.00
The percent susceptibility for trovafloxacin, tobramcyin, cefepime, and piperacillin using broth microdilution MICs and recent interpretative standards was 50%, 50%, 40%, and 40%, respectively (National Committee for Clinical Laboratory Standards 1999, Pfizer 1998). The overall agreement between broth microdilution and Etest MICs, allowing a two-fold dilution error, was 93%. In the three instances of disagreement, two of the disagreements were associated with a higher broth MIC than Etest MIC. Two of these three instances involved the same organism and both with -lactam antibiotics. Tobramycin 0.25⫻ MIC and 2⫻ MIC in combination with piperacillin 2⫻ MIC from the time-kill test was excluded from the analysis for all 10 isolates due to inactivation of tobramycin by piperacillin. The inactivation of tobramycin by piperacillin was demonstrated in repeat experiments and subsequent assay using fluorescence polarization immunoassay (AxSYM, Abbott Laboratories, Abbott Park, IL) to verify the actual tobramycin concentration. Using samples obtained from the time-kill at the end of 24 h, nearly complete inactivation of tobramycin was demonstrated when combined with piperacillin 2⫻ MIC (data not shown). Although inactivation likely occurred in the checkerboard test at these same concentrations, there was no apparent impact on the measured endpoint (turbidity), which is a less precise marker of effect. Similarly, inactivation may have occurred with the Etest, however, this was not assessed in this study. The FIC index and categorical interpretation by drug combination and multiple of the MIC are shown in Table 2.
J.A.Bosso et al. / Diagnostic Microbiology and Infectious Disease 38 (2000) 43–50
47
Table 2 FIC index and interpretation by drug combination Isolate
#1
#2
#3
#4
#5
#6
#7
#8
#9
#10
Drug Combination
Checkerboard
Etest
Time-Kill
Drug A/Drug B
FIC Index
FIC Index
0.25 ⫻ MIC A ⫹ 0.25 ⫻ MIC B
0.25 ⫻ MIC A ⫹ 2 ⫻ MIC B
2 ⫻ MIC A ⫹ 0.25 ⫻ MIC B
2 ⫻ MIC A ⫹ 2 ⫻ MIC B
TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI TV ⫹ CM TV ⫹ PI TO ⫹ CM TO ⫹ PI
NAa NAa 1.4 (A/I) 1.2 (A/I) 0.6 (A/I) 0.4 (S) 0.8 (A/I) 0.7 (A/I) 0.2 (S) 0.2 (S) 0.3 (S) 0.2 (S) 0.4 (S) 0.7 (A/I) 0.7 (A/I) 0.7 (A/I) 0.8 (A/I) 0.6 (A/I) 0.4 (S) 0.5 (A/I) NAa NAa 0.2 (S) 0.4 (S) NAa NAa 1.4 (A/I) 1.1 (A/I) NAa NAa 0.8 (A/I) 0.5 (S) 0.4 (S) 0.6 (A/I) 0.8 (A/I) 1.1 (A/I) NAa NAa 1.1 (A/I) 0.5 (S)
1.0 (A/I) 1.1 (A/I) 1.3 (A/I) 1.3 (A/I) 0.7 (A/I) 0.9 (A/I) 1.3 (A/I) 1.0 (A/I) 0.9 (A/I) 1.0 (A/I) 1.3 (A/I) 0.9 (A/I) 1.0 (A/I) 0.6 (A/I) 1.0 (A/I) 1.0 (A/I) 0.7 (A/I) 0.8 (A/I) 1.0 (A/I) 1.0 (A/I) 1.0 (A/I) 0.9 (A/I) 0.9 (A/I) 1.5 (A/I) 1.0 (A/I) 0.9 (A/I) 1.3 (A/I) 1.4 (A/I) 1.2 (A/I) 2.0 (A/I) 1.4 (A/I) 1.3 (A/I) 0.9 (A/I) 0.7 (A/I) 1.0 (A/I) 1.0 (A/I) 1.6 (A/I) 2.0 (A/I) 1.3 (A/I) 1.2 (A/I)
NAa NAa A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I NAa NAa A/I A/I NAa NAa A/I A/I NAa NAa A/I SYN A/I A/I A/I A/I NAa NAa A/I A/I
NAa NAa SYN NAb A/I SYN SYN NAb A/I A/I A/I NAb SYN SYN SYN NAb SYN A/I SYN NAb NAa NAa SYN NAb NAa NAa SYN NAb NAa NAa SYN NAb A/I A/I NAc NAb NAa NAa A/I NAb
NAa NAa A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I A/I SYN SYN A/I A/I NAa NAa SYN A/I NAa NAa SYN SYN NAa NAa SYN SYN ANT A/I SYN SYN NAa NAa A/I A/I
NAa NAa A/I NAb A/I A/I A/I NAb A/I A/I A/I NAb A/I A/I A/I NAb SYN A/I A/I NAb NAa NAa SYN NAb NAa NAa SYN NAb NAa NAa SYN NAb A/I A/I SYN NAb NAa NAa SYN NAb
NAa -Only includes isolates #2, 3, 4, 5, and 9 due to the insolubility of trovafloxacin in Mueller-Hinton broth NAb -Not included due to inactivation of tobramycin by piperacillin NAc -Not included due unavailability of data Abbreviations for interpretations: SYN, synergy; A/I, additivity/indifference; ANT, antagonism
Additivity/indifference was most common among all three methods. Antagonism was reported once with the time-kill test using trovafloxacin in combination with cefepime; detection of synergy, additivity/indifference, and antagonism varied depending upon antimicrobial combination and multiple of MIC. Synergy was detected most often when tobramycin was combined with cefepime (41%) and with combinations consisting of 0.25⫻ MIC of either trovafloxacin or tobramycin plus 2⫻ MIC of either cefepime or piperacillin (58%). With the checkerboard test, the FIC index ranged from 0.2 to 1.4. The incidence of synergy and ad-
ditivity/indifference was 37% and 63%, respectively. The FIC index of the Etest method varied from 0.6 to 2.0; neither synergy nor antagonism was detected. The agreement among the methods of synergy testing is shown in Table 3. Between the checkerboard and time-kill tests, the agreement was 51%. One major disagreement occurred using trovafloxacin and cefepime when synergy was reported with the checkerboard and antagonism was reported with the time-kill test. The remaining minor disagreements occurred when either additivity/indifference was reported with the checkerboard test and synergy was
48
J.A.Bosso et al. / Diagnostic Microbiology and Infectious Disease 38 (2000) 43–50
Table 3 Agreement between methods of synergy testing (%) Checkerboard Method
Etest Checkerboard
63 -
Time-Kill by Drugc
Time-Kill by Concentration 0.25 ⫻ MIC ⫹ 0.25 ⫻ MIC
0.25 ⫻ MIC ⫹ 2 ⫻ MICa
2 ⫻ MIC ⫹ 0.25 ⫻ MICb
2 ⫻ MIC ⫹ 2 ⫻ MIC
TV ⫹ CM
TV ⫹ PI
TO ⫹ CM
TO ⫹ PI
97 67
42 37
67 47
70 45
75 30
85 55
59 54
80 60
Combination of 0.25 ⫻ MIC of TV or TM and 2 ⫻ MIC of CF or PI Combination of 2 ⫻ MIC of TV or TM and 0.25 ⫻ MIC of CF or PI c Abbreviations for antibiotics: CM, cefepime; PI, piperacillin; TM, tobramycin; TV, trovafloxacin a
b
reported with the time-kill test or synergy was reported with the checkerboard test and additivity/indifference was reported with the time-kill test. Between the Etest and checkerboard, the agreement was 63%. No major disagreements occurred with these 2 methods and all minor disagreements occurred when additivity/indifference was reported with the Etest method and synergy was reported with the checkerboard. Between the Etest method and time-kill test, the agreement was 72%. Again, no major disagreements occurred and most minor disagreements occurred when additivity/indifference was reported with the Etest method and synergy was reported with time-kill test. Antagonism occurred once with the time-kill test when additivity/indifference was reported with the Etest method.
4. Discussion Acinetobacter baumannii is an organism that is increasingly becoming resistant to antimicrobials. Isolates resistant to broad spectrum antimicrobials have been identified, severely limiting the therapeutic choices available to clinicians and reinforcing the need to identify synergistic drug combinations (Afzal-Shah, 1998, Bergogne-Be´re´zin, 1996, Pascual, 1997, Tankovic, 1994, Weinbren, 1998). A technique that is simple to perform, reproducible, and detects the presence of in vitro synergy would be an important contribution to synergy detection methods and enhance its possible use in the clinical setting. We assessed synergy against 10 clinical isolates of Acinetobacter baumannii using the checkerboard and time-kill tests with trovafloxacin and tobramycin combined with cefepime and piperacillin. Synergy using the checkerboard test ranged from 20% with tobramycin and cefepime to 60% with trovafloxacin and cefepime. Overall, synergy and additivity/indifference was detected 37% and 63% of the time, respectively. These findings are similar to other investigators reporting synergy against susceptible and resistant strains of Acinetobacter spp. using the checkerboard test, ranging from 0% to 44% for each drug combination with antagonism rarely detected (Bajaksouzian, 1997, Chow, 1988, Martinez-Martinez, 1996). In the present study using the time-kill test, synergy was detected 27% of the time and ranged from 15% using trovafloxacin plus piperacillin to
41% using tobramycin plus cefepime. Only one occurrence of antagonism was noted and accounted for 1% of the results. Similarly, other investigators have detected synergy ranging from 0% to 93% for a combination of a -lactam and either an aminoglycoside or a fluoroquinolone using the time-kill test against Acinetobacter spp. (Bajaksouzian, 1997, Glew, 1977). We evaluated a new method to assess synergy using Etest strips and compared it to the checkerboard and timekill tests. The agreement between the Etest method and checkerboard test (63%) and between the Etest method and time-kill test (72%) was higher than between the checkerboard and time-kill tests (51%). When disagreement occurred between the Etest method and either checkerboard or time-kill test, the Etest method was generally more conservative (e.g., when synergy was observed using either the checkerboard or time-kill tests, additivity/indifference was observed using the Etest method). These results are similar to a previously published comparison of the checkerboard, time-kill, and Etest methods (White, 1996). In that study, the agreement between the Etest method and checkerboard test was 75%, whereas the agreement between the Etest method and time-kill test ranged from 63% to 75%. Additionally, antagonism was not detected using the Etest method in that study. Furthermore, when antagonism was observed using the time-kill test in that study, additivity/ indifference was reported using the Etest method. As previously discussed, antagonism may be difficult to detect with the Etest method and this may explain discordance when antagonism was found using the other methods (White, 1996). Synergy was detected with both the checkerboard and time-kill tests using isolates not susceptible to one or both antimicrobials based on current interpretive standards (National Committee for Clinical Laboratory Standards 1999, Pfizer 1998). Using the checkerboard test, four of 11 instances of synergy occurred when the isolate was intermediate or resistant to one antimicrobial, whereas four of 11 instances of synergy occurred when the isolate was intermediate or resistant to both antimicrobials. With the timekill test, synergy occurred in one of 27 instances when the isolate was intermediate to one antimicrobial and in 14 of 27 instances when the isolate was intermediate or resistant to both antimicrobials. Other researchers have also observed
J.A.Bosso et al. / Diagnostic Microbiology and Infectious Disease 38 (2000) 43–50
synergy with isolates resistant to a -lactam and/or aminoglycoside using the time-kill test with drug concentrations ranging from 0.25⫻ MIC to 4⫻ MIC (Cappelletty, 1996). One may question the relevance of additivity/indifference or synergy if the MIC of one or both agents is not decreased below the susceptibility breakpoint. In this study, a decrease in the MIC of an antimicrobial agent resulting in a categorical change from intermediate or resistant to susceptible was observed using the checkerboard test and Etest methods. Although this type of change was observed in nine different drug/isolate combinations with the checkerboard test, in only 67% of these was synergy detected. Using the Etest method, this type of categorical susceptibility change occurred in five different drug/isolate combinations. All of these combinations were found to be additive/indifferent. There were no instances with either the checkerboard test or Etest method in which a susceptibility categorical change (e.g., resistant to susceptible) to both antimicrobial agents resulted. In most instances in which the susceptibility category changed, the initial susceptibility interpretation was intermediate and the MIC never decreased more than fourfold. Therefore, antimicrobial combinations that result in a lowering of the MIC and ultimately a change in the susceptibility category may not directly correlate with achieving synergy. Although one might expect that combinations containing 2⫻ MIC of each agent would result in the greatest frequency of synergy, this was not observed in our study. Since the definition of synergy requires a 2 log increase in killing by the combination compared to the single most active agent, it is difficult to detect synergy when single agents exhibit extensive killing. This was illustrated in our study by combinations containing trovafloxacin or tobramycin at 2⫻ MIC. In 60% of the time-kill tests with this multiple of the MIC of these agents, final colony counts with the single agents were ⱕ 1.03 ⫻ 103 CFU/mL and synergy could not be detected based on our limit of quantitation. Had we excluded those combinations where the activity of a single drug prevented them from meeting the definition of synergy, the rate of synergy and the agreement between the time-kill and Etest or time-kill and checkerboard methods would have varied from the results cited above. Further, the highest rate of agreement would have been observed between the Etest and checkerboard (63%), followed by the Etest and time-kill (39%) and time-kill and checkerboard (30%). Conversely, one might expect with combinations containing 0.25⫻ MIC of each agent synergy might be more easily detected if it occurs. However, combinations containing these multiples of the MIC minimally impacted the killing over the more active single agent in addition to being unlikely to result in sufficient killing of the initial inoculum to meet the criterion of synergy. As an additional observation related to the definition of synergy, we noted that for all drug combinations studied, the least active drug produced a ⬍100-fold reduction in bacteria at 24 h in relation to the growth control curve.
49
In this study with Acinetobacter baumannii, we detected varying degrees of synergy with all drug combinations with one or more methods. The clinical utility of such combinations requires evaluation. Among the methods evaluated in this study, we found the highest rate of agreement between the Etest and the time-kill test. However, when drug combinations in which the activity of one of the drugs prevented the combination from meeting the definition of synergy were excluded, the highest rate of agreement was associated with the Etest and checkerboard methods. While one could debate the appropriateness of comparing test results from methods which rely on different endpoints, the checkerboard and time-kill methods represent the current accepted standard for such studies and thus serve as a point of reference for any newer method. It is also recognized that it is possible to garner far greater information from a time-kill test than that represented by observations at 0 and 24 h only. Colony counts at additional time points may reflect more fully the nature of the drug-drug and drug(s)-organism interaction. Nonetheless, the standard method for use of this test in synergy testing relies solely on 0 and 24 h observations, as used in the present study. As far as the new method is concerned, the Etest method is less labor intensive and easily performed, although it appears to be more conservative than either the checkerboard or time-kill tests in detecting synergy. While this new method of synergy testing appears promising, further work must be performed using additional organisms and antimicrobial combinations before it is more widely employed, especially in the face of the fact that no synergy was detected with Etest, even when present according to time-kill and/or checkerboard. Correlation of these methods with outcomes from clinical studies is needed to determine the ideal method of synergy testing. Acknowledgments This work was supported, in part, by a grant from Pfizer Inc., New York, NY. We acknowledge Linda B. Mihm, Pharm.D. for her contribution to the study. References Afzal-Shah, M., & Livermore D. M. (1998). Worldwide emergence of carbapenem-resistant Acinetobacter spp. J Antimicrob Chemother, 41, 576 –577. Bajaksouzian, S., et al. (1997). Activities of levofloxacin, ofloxacin, and ciprofloxacin, alone and in combination with amikacin, against Acinetobacters as determined by checkerboard and time-kill studies. Antimicrob Agents Chemother, 41, 1073–1076. Bergogne-Be´re´zin, E., et al. (1987). Epidemiology of nosocomial infections due to Acinetobacter calcoaceticus. J Hosp Infect, 10, 105–113. Bergogne-Be´re´zin, E., & Joly-Guillou, M. L. (1991). Hospital infection with Acinetobacter spp.: an increasing problem. J Hosp Infect, 18(Suppl A), 250 –255. Bergogne-Be´re´zin, E., & Towner, K. J. (1996). Acinetobacter spp. as nosocomial pathogens: microbiological, clinical, and epidemiological features. Clin Microbiol Rev, 9, 148 –165.
50
J.A.Bosso et al. / Diagnostic Microbiology and Infectious Disease 38 (2000) 43–50
Bouvet, P. J. M., & Grimont P. A. D. (1987). Identification and biotyping of clinical isolates of. Acinetobacter Ann Inst Pasteur Microbiol, 138, 569 –578. Cappelletty, D. M., Rybak, M. J. (1996). Comparison of methodologies for synergism testing of drug combinations against resistant strains of Pseudomonas aeruginosa. Antimicrob Agents Chemother, 40, 677– 683. Chow, A. W., et al. (1988). Synergistic interactions of ciprofloxacin and extended spectrum -lactams or aminoglycosides against Acinetobacter calcoaceticus ss. anitratus. Diagn Microbiol Infect Dis, 9, 213– 217. Eliopoulos, G. M., & Moellering, R. C. (1996). Antimicrobial combinations. In Antibiotics in Laboratory Medicine 4th ed, Ed V. Lorian. Baltimore: Williams & Wilkins, pp. 330 –396. Glew, R. H., et al. (1977). In vitro synergism between carbenicillin and aminoglycosidic aminocyclitols against Acinetobacter calcoaceticus var. anitratus. Antimicrob Agents Chemother, 6, 1036 –1041. Martinez-Martinez , L., et al. (1996). In-vitro activity of antimicrobial agent combinations against multiresistant Acinetobacter baumannii. J Antimicrob Chemother, 38, 1107–1108. National Committee for Clinical Laboratory Standards. (1993). Methods for dilution antimicroibal susceptibility tests for bacteria that grow
aerobically: NCCLS Document M7–A3, Vol. 13, No. 25 National Committee for Clinical Laboratory Standards, Villanova, PA. National Committee for Clinical Laboratory Standards. (1999). Performance standards for antimicrobial susceptibililty testing: NCCLS Document M100 –S9, Vol. 19, No. 1 National Committee for Clinical Laboratory Standards, Wayne, PA. Ng, P. C., et al. (1989). An outbreak of Acinetobacter septicaemia in a neonatal intensive care unit. J Hosp Infect, 14, 363–368. Pascual, A. (1997). In-vitro susceptibilies of multiresistant strains of Acinetobacter baumannii to eight quinolones. J Antimicrob Chemother, 40, 140 –142. Pfizer, Inc. Trovan package insert. New York, NY; December 1998. Tankovic, J., et al. (1994). Characterization of a hospital outbreak of imipenem-resistant Acinetobacter baumannii by phenotypic and genotypic typing methods. J Clin Microbiol, 32, 2677–2681. Weinbren, M. J., et al. (1998). Acinetobacter spp. isolates with reduced susceptibilities to carbapenems in a UK burns unit. J Antimicrob Chemother, 41, 574 –576. White, R. L., et al. (1996). Comparison of three different in vitro methods of detecting synergy: Time-kill, checkerboard, and Etest. Antimicrob Agents Chemother, 40, 1914 –1918.