Extensively-resistant Acinetobacter baumannii (XDR-AB) presents an enormous challenge to health care, particularly in intensive care units (ICU)   . The presence of strains resistant to all available antibiotics has led to reliance on the polymyxins as the last resort. This group of antibiotics was used in the 1950’s, but due to their neurotoxicity and nephrotoxicity, there was a massive decline in their use  .
Another problem with colistin especially with multidrug resistant isolates is heteroresistance that has been observed in vitro and has also developed during therapy   , raising issues about colistin being used alone as a monotherapy which may lack sufficient killing activity and aid in selection of resistant subpopulation  .
Polymyxins work through disrupting the integrity of the Gram-negative bacterial membrane, increasing its permeability to substances that are usually excluded, thus increasing the activity of hydrophobic antibiotics, which otherwise had no effect    . One of these antibiotics is the glycopeptide vancomycin, where due to its large size and hydrophobicity; it lacks the power to exert its action against Gram-negative bacilli. Using colistin could improve the penetration of glycopeptides through bacterial membrane. Also azythromycin, the most commonly prescribed antibiotic in the U.S., is never recommended for clinical treatment of serious Gram negative infections because of poor or absent in vitro activity by standard in-vitro testing. However, antibacterial activity of azithromycin was found to be enhanced in tissue culture media vs. bacteriologic media, prompting closer examination of its interaction with drug resistant Gram negative bacilli   .
In order to explore the potential usefulness of such antibiotics, we attempted to evaluate the presence of in-vitro activity (synergy) of colistin in combination with vancomycin or azithromycin against XDR-AB, in an attempt to find out the possibility of such combination therapy, to minimize the toxic effects and to prevent the development of resistance when using colistin alone.
2. Material & Methods
2.1. Clinical Isolates
Thirty XDR-AB clinical isolates were included in this study. Isolates were collected from urinary, blood, pus and respiratory samples received at the diagnostic Microbiology Lab of Alexandria Main University Hospital over a period of one year starting January till December 2015. Acinetobacter initial identification was done via conventional biochemical methods (oxidase negative, citrate positive, non-motile and triple sugar iron agar negative) and confirmed to the species level by (Vitek 2 compact, bioMérieux, France)   .
XDR was defined according to European Centre for Disease Prevention and Control (ECDC) and the Centers for Disease Control and Prevention (CDC) as non-susceptibility to at least one agent in all but two or fewer antimicrobial categories (i.e. bacterial isolates remain susceptible to only one or two categories). Antibiotic susceptibility was interpreted according to CLSI recommendations  .
2.2. Antimicrobial Susceptibility Testing of A. baumannii Isolates
2.2.1. Automated Vitek 2 Compact System
The Vitek 2 susceptibility card (GN222 AST card) (bioMérieux, France) including a colistin susceptibility test was used. Interpretive breakpoints (MIC ≤ 2 μg/ml, susceptible, and MIC ≥ 4 μg/ml, resistant) were considered for the Vitek 2 susceptibility testing according to the manufacturer's instructions. A. baumannii ATCC 19606 was used as a control strain.
The GN222 AST card tests susceptibility against the following antibiotics besides colistin (CST): ticarcillin (TIC), ticarcillin/clavulanic acid (TIC/CA), piperacillin (PIP), piperacillin/tazobactam (TZP), ceftazidime (CAZ), cefepime (FEP), Aztreonem (ATM), imipenem (IMP), meropenem (MEM), amikacin (AMK), gentamicin (CN), tobramycin (TOB), ciprofloxacin (CIP), pefloxacin (PEF), minocycline (MN), rifampicin (RP) and trimethoprim/sulphamethoxazole (SXT).
2.2.2. Disc Diffusion Test
To cover the panel of antibiotics suggested by CLSI, disc diffusion was used for ampicillin/sulbactam (SAM), levofloxacin (LVX), tetracycline (TC), doxycycline (DC), cefotaxime (CTX) and ceftriaxone (CRO) according to CLSI 2015 guidelines. As for tigecycline (TGC) breakpoints suggested by Piewngam et al. were used in disc diffusion testing where a zone diameter of ≥17 mm was considered sensitive, while a diameter of ≤12 mm was considered resistant  .
Besides testing colistin susceptibility by Vitek 2 compact system, the disc diffusion method using 10 μg disc was also done (≤ 12 mm = resistant and ≥ 14 mm = susceptible)  . Since neither the current CLSI nor EUCAST provided disc diffusion zone diameter breakpoints for colistin against Acinetobacter spp., and since both recommended that colistin testing should be performed by dilution method and interpreted according to MIC, colistin susceptibility disc diffusion results were compared against broth microdilution method (as a reference method)   .
Colistin Categorical agreement (CA) was defined as the percentage of isolates classified in the same susceptibility category by broth microdilution method and the disc diffusion or Vitek. Very major errors (VMEs) denoted a false-susceptible result, and major errors (MEs) denoted a false-resistant result, while minor errors (MinE) were intermediate zone diameters that had susceptible or resistant MIC, or intermediate MIC with a susceptible or resistant zone diameter. Acceptable performance was evaluated according to criteria established by the International Organization for Standardization: ≥90% for category agreement and ≤3% for VMEs or MEs  .
2.3. Checkerboard Synergy Test
The checkerboard method was performed according to the method described by Schwalbe (2007)  .
2.3.1. Preparation of Checkerboard Antibiotic Microdilution
Antibiotic powders used were colistin sulfate salt, vancomycin, and azithromycin (Sigma-Aldrich, USA). The concentration of antibiotic prepared prior to dilution was four times higher than the highest concentration to be tested. For example, if the initial concentration for antibiotic A was 256 μg/ml and for antibiotic B was 8 μg/ml, then it was necessary to start with a concentration of 1024 μg/ml for antibiotic A and 32 μg/ml for antibiotic B.
Checkerboards were prepared by doubling dilutions of vancomycin, or azithromycin (0 to 256 μg/ml) in the horizontal wells and colistin sulfate (0 to 8 μg/ml) in the vertical wells using the format shown in Table 1.
2.3.2. Preparation of Inoculum
The inoculum was prepared using a 1:100 dilution of a half McFarland bacterial suspension from an overnight culture (approximately 5 × 105 CFU/ml of bacteria).
2.3.3. Inoculation and Incubation of Checkerboard Panel
50 μl was inoculated into each well of a 96-well microtiter plate. The plates were covered and incubated at 35˚C for 18 - 24 h.
2.3.4. Reading the Results
1) The wells without visible signs of growth were identified visually against a dark background. Then the MICs (lowest concentration showing inhibition of growth) for the individual antibiotics in the checkerboard method were recorded.
2) The lowest fractional inhibitory concentration index (FICI) was used to define synergy.
2.3.5. Data Analysis
FICI was used to analyze data from the checkerboard assay. The FICI was calculated using the following equation: the FIC of antibiotic = MIC of antibiotic in combination/MIC of antibiotic alone and FICI = FIC of colistin + FIC of Vancomycin or azithromycin. The results of the FICI was interpreted as follows: synergy, FICI ≤ 0.5; antagonism, FICI > 4; indifference, 0.5 < FICI ≤ 4  .
Table 1. The checkerboard format used in the study to test in-vitro synergy of colistin and vancomycin/azithromycin (where colistin dilutions were distributed horizontally in rows (B to H) while Vancomycin or azithromycin dilutions were distributed vertically in columns (2-12).
The 30 XDR-AB were isolated from respiratory; broncho-alveolar lavage and sputum (20/30), Pus (6/30), Urine (2/30), CSF (1/30) and Blood (1/30) samples delivered to AMUH microbiology lab. Intensive care units being the source of 73.3% (22/30) of A. baumannii isolates in our study.
3.1. Susceptibility Testing Results
The susceptibility patterns of the strains isolated are shown in Table 2. Twenty eight isolates were sensitive to colistin MIC ≤ 2 μg/ml, only 2 were found resistant with MIC ≥ 4 μg/ml by broth microdilution method (reference method) and by the Vitek 2 system, and these two strains were pan-resistant, isolated at different periods during the study. However, disc diffusion failed to detect these 2 resistant strains, thus identified 100% of the isolates as susceptible.
Table 2. Distribution of antimicrobial profile of the 30 XDR A. baumannii strains.
Table 3. Comparison of colistin susceptibility by disc diffusion and broth microdilution test for the 30 XDR A. baumannii strains.
On comparing the two methods used for detection of susceptibility to colistin to broth microdilution MIC reference method, the Vitek showed 100% CA with no errors among the 30 isolates tested, on the other hand the disc diffusion although showed CA = 93%, it failed to detect the 2 resistant strains thus showing very major errors (Table 3).
3.2. Checkerboard Synergy Test
The checkerboard synergy test was performed for each isolated A. baumanniii strain against, colistin/vancomycin and colistin/azythromycin combinations. All strains were found resistant to vancomycin and azithromycin (mic ≥ 256 μg/ml) by broth microdilution. However, when colistin was added in the checkerboard format and the lowest (FICI) was used to analyse the outcome, synergy was detected for all isolates (100%) with vancomycin (FICI ranged from 0.03 to 0.0241 i.e. ≤0.5), mean FICI = 0.08. As for azithromycin 21 strains had FICI range from 0.7 to 1.001 denoting indifference; 0.5 < FICI ≤ 4, the remaining 9 strains showed synergy with FICI range from 0.06 to 0.241. The mean colistin/azithromycin FICI = 0.71 for the 30 isolates.
When combining colistin and vancomycin, there was a reduction in colistin MIC level from 8 μg/ml to ≤0.5 μg/ml for all isolates, where 100% of the isolates showed decreased MIC level below the susceptible breakpoints (≤2 μg/ml) of them 80% (24/30) showed an MIC level of 0.12 μg/ml after combination. As for vancomycin, all isolates (100%) showed reduction in MIC level from ≥256 μg/ml to 0.25 μg/ml.
On the other hand, the colistin MIC levels showed a decrease from 8 μg/ml to ≤2 μg/ml for all isolates when combined to azithromycin; 80% (24/30) of the isolates’ colistin MICs were ≤1 μg/ml. The azithromycin MIC decreased from ≥256 μg/ml to 0.25 μg/ml for all isolates.
Acinetobacter drug resistance is increasing, reducing the treatment options available for managing such infections. A. baumannii are considered as major infectious threat especially in intensive care units (ICU), where it has been implemented in the treat- ment of various nosocomial infections especially pneumonia, urinary tract infections and bloodstream infections   . Our study has similar findings to these data, ensuring that acinetobacter ICU infections are a global trend. Also respiratory specimens had the highest rate of A. baumanniii isolation, constituting (20/30) 66.7% of all specimens, similar findings were found by Al Bshabshe et al. 2016  .
The last resort for treatment of drug resistant A. baumannii is colistin where it is often the only agent with in vitro activity as shown in our results; 28/30 XDR-AB strains were found sensitive to colistin, however, clinical experience of its use for the treatment of resistant strains as a monotherapy has not always been successful resulting in poor outcomes  . However, the optimal method for testing susceptibility of A. baumannii to colistin is still controversial. The drug diffusion through the agar is inconsistent and poor, making errors with disc diffusion common when compared to other methods  . The controversy comes from the absence of guidelines by the CLSI for disc diffusion. During our work we got to study three colistin susceptibility testing methods against A. baumannii clinical isolates with XDR patterns  . In the current study the disc diffusion although showing CA of 93% yet it failed to detect the two resistant A. baumanniii detected by the broth microdilution (reference method) giving VMEs, however, basing our findings on only these resistant strains is inconclusive but in the absence of verified guidelines, disc diffusion method should not be used alone in judging colistin susceptibility in A. baumanniii. Our finding is supported by others who reported disc diffusion to be an unreliable method for detecting polymyxins susceptibility   . As regards Vitek 2 AST, we found it to be reliable and easy with 100% CA, Lee et al. 2013 tested 213 isolates, including 13 colistin resistant Acinetobacter strains. Vitek 2 showed excellent CA with 0.9% VMEs and no ME  .
In addition, another problem is that trends are showing elevated colistin MICs globally, emphasizing not only the importance of accurate colistin susceptibility testing, but also the importance of appropriate combination therapy to prevent the emergence of resistance to colistin by achieving synergy  . Also when a clinical strain is resistant to all antibiotic the achievement of synergy might not be a reachable option by combination therapy, yet any antibiotic combination activity would be preferable to the inactivity of a single drug. Thus an additive or subadditive effect would be welcomed. This can also be achieved when having a single active agent boosted by an inactive agent. In this case, prevention of resistance to the active agent may be possible  .
Our in vitro study demonstrated the possibility for a vancomycin-colistin combination, where synergy was demonstrated for all isolates tested, however, concern exists regarding its clinical application. But Gordon et al. hypothesized, that since synergistic combinations decreased the dose of colistin required to inhibit Acinetobacter, it may be possible to produce synergy in vivo by using lower-than-normal doses of colistin, a strategy similar to the use of low-dose aminoglycosides in combination with β-lactams in the treatment of streptococcal endocarditis  . They also supported the explanation for the mechanism behind this synergy by electron microscopy imaging that revealed disruption of membranes of colistin exposed bacteria in comparison to those of unexposed controls.
Regarding testing combination of colistin and azithromycin using the checkerboard method and although azithromycin on its own shows negligible effect against MDR- GNRs. Yet, we found that when combined with colistin it did exhibit an activity against drug resistant acinetobacter although this combination only produced a synergistic effect on 30% of our isolates, similarly a small number of studies demonstrated the same activity for AZM against Gram negative bacilli with no explanation or suggestion about the mechanism of synergy   . Until Lin et al. 2012 tested this combination in eukaryotic cell media and in vivo murine models of infection and concluded that AZM entry and activity is synergistically enhanced when the bacterial outer membrane was disrupted by colistin  . Buyck et al., 2012 demonstrated that a mutation of the oprM efflux pump system in P. aeruginosa lead to an increased AZM susceptibility, and on entry AZM reduced oprM gene expression and protein synthesis, further enhancing the entry of AZM in eukaryotic media and in synergy with colistin that may initiate a positive feedback loop to increase effective intracellular levels of the antibiotic  .
In conclusion, this study opens the door to further explore antibiotics, not known to act on Gram negative bacilli as vancomycin and azythromicin for adjunctive therapy in MDR infections. Potentially, other antibiotics not considered due to large molecular size deterring penetration through Gram-negative bacterial membranes in standard MIC testing, should be considered with colistin, allowing a lower dose and reducing the side effects. The findings of the present study suggest that regimens containing vancomycin may confer therapeutic benefit for infection due to XDR A. baumannii. However, other methods (time-kill assay) should be used to confirm such synergy. Furthermore, the optimal combination treatment for serious XDR-AB infection should be addressed in a prospective clinical trial.
 Coelho, J.M., Turton, J.F., Kaufmann, M.E., Glover, J., Woodford, N., Warner, M., Palepou, M.F., Pike, R., Pitt, T.L., Patel, B.C. and Livermore, D.M. (2006) Occurrence of Carbapenem-Resistant Acinetobacter baumannii Clones at Multiple Hospitals in London and Southeast England. Journal of Clinical Microbiology, 44, 3623-3627. https://doi.org/10.1128/JCM.00699-06
 Falagas, M.E. and Kasiakou, S.K. (2005) Colistin: The Revival of the Polymyxins for the Management of Multi-Drug Resistant Gram-Negative Bacterial Infections. Clinical Infectious Diseases, 40, 1333-1341. https://doi.org/10.1086/429323
 Hawley, J.S., Murray, C.K. and Jorgensen, J.H. (2008) Colistin Heteroresistance in Acinetobacter and Its Association with Previous Colistin Therapy. Antimicrobial Agents and Chemotherapy, 52, 351-352. https://doi.org/10.1128/AAC.00766-07
 Ko, K.S., Suh, J.Y., Kwon, K.T., Jung, S.I., Park, K.H., Kang, C.I., Chung, D.R., Peck, K.R. and Song, J.H. (2007) High Rates of Resistance to Colistin and Polymyxin B in Subgroups of Acinetobacter baumannii Isolates from Korea. Journal of Antimicrobial Chemotherapy, 60, 1163-1167. https://doi.org/10.1093/jac/dkm305
 Pachon-Ibanez, M.E., Docobo-Perez, F., Lopez-Rojas, R., Domínguez-Herrera, J., Jimenez-Mejias, M.E., García-Curiel, A., Pichardo, C., Jiménez, L. and Pachon, J. ( 2010) Efficacy of Rifampin and Its Combinations with Imipenem, Sulbactam, and Colistin in Experimental Models of Infection Caused by Imipenem-Resistant Acinetobacter baumannii. Antimicrobial Agents and Chemotherapy, 54, 1165-1172. https://doi.org/10.1128/AAC.00367-09
 Li, J., Nation, R.L., Milne, R.W., Turnidge, J.D. and Coulthard, K. (2005) Evaluation of Colistin as an Agent against Multi-Resistant Gram-Negative Bacteria. International Journal of Antimicrobial Agents, 25, 11-25. https://doi.org/10.1016/j.ijantimicag.2004.10.001
 Lam, C., Hildebrandt, J., Schutze, E. and Wenzel, A. (1986) Membrane Disorganizing Property of Polymyxin B Nonapeptide. Journal of Antimicrobial Chemotherapy, 18, 9-15.
 Buyck, J.M., Plesiat, P., Traore, H., Vanderbist, F., Tulkens, P.M. and Van Bambeke, F. (2012) Increased Susceptibility of Pseudomonas aeruginosa to Macrolides and Ketolides in Eukaryotic Cell Culture Media and Biological Fluids due to Decreased Expression of oprM and Increased Outer-Membrane Permeability. Clinical Infectious Diseases, 55, 534-542.
 Lin, L., Nonejuie, P., Munguia, J., Hollands, A., Olson, J., Dam, Q., et al. (2015) Azithromycin Synergizes with Cationic Antimicrobial Peptides to Exert Bactericidal and Therapeutic Activity against Highly Multidrug-Resistant Gram-Negative Bacterial Pathogens. EBioMedicine, 2, 690-698. https://doi.org/10.1016/j.ebiom.2015.05.021
 Constantiniu, S., Romaniuc, A., Iancu, L.S., Filimon, R. and Taraşi, I. (2004) Cultural and Biochemical Characteristics of Acinetobacter spp. Strains Isolated from Hospital Units. Journal of Preventive Medicine, 12, 35-42.
 Zbinden, A., Böttger, E.C., Bosshard, P.P. and Zbinden, R. (2007) Evaluation of the Colorimetric VITEK 2 Card for Identification of Gram-Negative Nonfermentative Rods: Comparison to 16S rRNA Gene Sequencing. Journal of Clinical Microbiology, 45, 2270-2273.
 Magiorakos, A.-P., Srinivasan, A., Carey, R.B., Carmeli, Y., Falagas, M.E., Giske, C.G., Harbarth, S., Hindler, J.F., Kahlmeter, G., Olson-Liljequist, B., et al. (2012) Multidrug-Resistant, Extensively Drug-Resistant and Pandrug-Resistant Bacteria: An International Expert Proposal for Interim Standard Definitions for Acquired Resistance. Clinical Microbiology and Infections, 18, 268-281. https://doi.org/10.1111/j.1469-0691.2011.03570.x
 Piewngam, P. and Kiratisin, P. (2014) Comparative Assessment of Antimicrobial Susceptibility Testing for Tigecycline and Colistin against Acinetobacter baumannii Clinical Isolates, Including Multidrug-Resistant Isolates. International Journal of Antimicrobial Agents, 44, 396-401.
 Galani, I., Kontopidou, F., Souli, M., Rekatsina, P., Koratzanis, E., Deliolanis, J., et al. (2008) Colistin Susceptibility Testing by Etest and Disk Diffusion Methods. International Journal of Antimicrobial Agents, 31, 434-439. https://doi.org/10.1016/j.ijantimicag.2008.01.011
 Clinical and Laboratory Standards Institute (2015) Performance Standards for Antimicrobial Susceptibility Testing; 25th Informational Supplement. CLSI Document M100-S25, Clinical and Laboratory Standards Institute, Wayne, PA.
 Clinical and Laboratory Standards Institute (2015) Methods for Dilution of Antimicrobial Susceptibility Tests for Bacteria That Grow Aerobically; Approved Standard—10th Edition. CLSI Document M07-A10, Clinical and Laboratory Standards Institute, Wayne, PA.
 International Organization for Standardization (2007) ISO 20776-2:2007(E): Clinical Laboratory Testing and in Vitro Diagnostic Test Systems—Susceptibility Testing of Infectious Agents and Evaluation of Performance of Antimicrobial Susceptibility Test Devices—Part 2: Evaluation of Performance of Antimicrobial Susceptibility Test Devices. International Organization for Standardization, Geneva, 9 p.
 Petersen, P.J., Labthavikul, P., Jones, C.H. and Bradford, P.A. (2006) In Vitro Antibacterial Activities of Tigecycline in Combination with Other Antimicrobial Agents Determined by Chequerboard and Time-Kill Kinetic Analysis. Journal of Antimicrobial Chemotherapy, 57, 573-576. https://doi.org/10.1093/jac/dki477
 Dijkshoorn, L., Nemec, A. and Seifert, H. (2007) An Increasing Threat in Hospitals: Multidrug-Resistant Acinetobacter baumannii. Nature Reviews of Microbiology, 5, 939-951.
 Al Bshabshe, A., Joseph, M.R.P., Al Hussein, A., Haimour, W. and Hami, M. (2016) Multidrug Resistance Acinetobacter Species at the Intensive Care Unit, Aseer Central Hospital, Saudi Arabia: A One Year Analysis. Asian Pacific Journal of Tropical Medicine, 9, 903-908.
 Livermore, D.M., Hill, R.L., Thomson, H., Charlett, A., Turton, J.F., Pike, R., Patel, B.C., Manuel, R., Gillespie, S., Balakrishnan, I., Barrett, S.P., Cumberland, N., Twagira, M. and the C-MRAB Study Group (2010) Antimicrobial Treatment and Clinical Outcome for Infections with Carbapenem- and Multiply-Resistant Acinetobacter baumannii around London. International Journal of Antimicrobial Agents, 35, 19-24. https://doi.org/10.1016/j.ijantimicag.2009.09.014
 Lo-Ten-Foe, J.R., De Smet, A.M., Diederen, B.M., Kluytmans, J.A. and Van Keulen, P.H. (2007) Comparative Evaluation of the VITEK 2, Disk Diffusion, Etest, Broth Microdilution, and Agar Dilution Susceptibility Testing Methods for Colistin in Clinical Isolates, Including Heteroresistant Enterobacter cloacae and Acinetobacter baumannii Strains. Antimicrobial Agents and Chemotherapy, 51, 3726-3730. https://doi.org/10.1128/AAC.01406-06
 European Centre for Disease Prevention and Control (2013) Antimicrobial Resistance Surveillance in Europe. Annual Report of the European Antimicrobial Resistance Surveillance Network (EARS-Net). European Centre for Disease Prevention and Control, Stockholm.
 Tan, T.Y. and Ng, L.S.Y. (2006) Comparison of Three Standardized Disc Susceptibility Testing Methods for Colistin. Journal of Antimicrobial Chemotherapy, 58, 864-867.
 Lee, S.Y., Shin, J.H., Lee, K., Joo, M.Y., Park, K.H., Shin, M.G., et al. (2013) Comparison of the Vitek 2, MicroScan, and Etest Methods with the Agar Dilution Method in Assessing Colistin Susceptibility of Bloodstream Isolates of Acinetobacter Species from a Korean University Hospital. Journal of Clinical Microbiology, 51, 1924-1926. https://doi.org/10.1128/JCM.00427-13
 Rahal, J.J. (2006) Novel Antibiotic Combinations against Infections with Almost Completely Resistant Pseudomonas aeruginosa and Acinetobacter Species. Clinical Infectious Diseases, 43, S95-S99. https://doi.org/10.1086/504486
 Gordon, N.C., Png, K. and Wareham, D.W. (2010) Potent Synergy and Sustained Bactericidal Activity of a Vancomycin-Colistin Combination versus Multidrug-Resistant Strains of Acinetobacter baumannii. Antimicrobial Agents and Chemotherapy, 54, 5316-5322.
 Timurkaynak, F., Can, F., Azap, O.K., Demirbilek, M., Arslan, H. and Karaman, S.O. (2006) In Vitro Activities of Non-Traditional Antimicrobials Alone or in Combination against Multidrug-Resistant Strains of Pseudomonas aeruginosa and Acinetobacter baumannii Isolated from Intensive Care Units. International Journal of Antimicrobial Agents, 27, 224-228.
 Vaara, M., Siikanen, O., Apajalahti, J., Fox, J., Frimodt-Moller, N., He, H., Poudyal, A., Li, J., Nation, R.L. and Vaara, T. (2010) A Novel Polymyxin Derivative That Lacks the Fatty Acid Tail and Carries Only Three Positive Charges Has Strong Synergism with Agents Excluded by the Intact Outer Membrane. Antimicrobial Agents and Chemotherapy, 54, 3341-3346.