Received 27 August 2002; returned 26 November 2002; revised 8 January 2003; accepted 11 January 2003

Journal of Antimicrobial Chemotherapy (2003) 51, 905 911 DOI: 10.1093/jac/dkg152 Advance Access publication 13 March 2003 AUC 0 t /MIC is a continuous index of fluoroquinolone exposure and predictive of antibacterial response for Streptococcus pneumoniae in an in vitro infection model Sheryl A. Zelenitsky 1 3 *, Robert E. Ariano 1 3, Harris Iacovides 1, Siyao Sun 1 and Godfrey K. M. Harding 2,4,5 Faculties of 1 Pharmacy and 2 Medicine, University of Manitoba; 3 Pharmacy, 4 Infectious Diseases and 5 Microbiology Laboratory, St Boniface General Hospital, Winnipeg, MB, Canada Received 27 August 2002; returned 26 November 2002; revised 8 January 2003; accepted 11 January 2003 Objective: To conduct a comprehensive pharmacodynamic analysis of moxifloxacin and levofloxacin against Streptococcus pneumoniae in an in vitro infection model. Methods: In dose escalation studies, single doses with peak concentrations equivalent to 1, 2, 4, 8, 16 and 32 MIC against two isolates of S. pneumoniae were studied over 24 h. Traditional pharmacodynamic indices, including peak concentration divided by MIC (peak/mic), time of concentration above MIC (T > MIC) and AUC 24 /MIC, were estimated for all regimens. As a continuous index of fluoroquinolone exposure, AUC 0 t /MIC was also calculated, as AUC from time 0 to 1, 2 and 6 h divided by MIC. Correlations between pharmacodynamic indices and antibacterial effects were examined using linear and non-linear methods. In validation experiments, the pharmacodynamic model was used to predict bacterial kill curves, produced by simulated clinical doses of moxifloxacin and levofloxacin against two other S. pneumoniae isolates. Results: Peak/MIC was most predictive of early bacterial kill, whereas T > MIC was significantly associated with final bacterial counts at 24 h. Antibacterial effects were bacteriostatic when T > MIC was 48% and bactericidal when values exceeded 55%. AUC 0 t /MIC was strongly associated with bacterial kill throughout the dosing interval. Bactericidal activity and bacterial eradication were associated with AUC 0 t /MICs of 28 and 135, respectively. AUC 0 t /MIC was also highly predictive of bacterial kill curves produced by simulated clinical doses of moxifloxacin and levofloxacin (precision 0.36 log 10 cfu/ml, bias 0.02 log 10 cfu/ml). Conclusion: This study demonstrated the novel application of AUC 0 t /MIC as a continuous index of antibiotic activity, and provided extensive characterization of fluoroquinolone pharmacodynamics against S. pneumoniae. Keywords: moxifloxacin, levofloxacin, pharmacodynamics, pneumococcus Introduction Pharmacodynamic (PD) research explores complex antibiotic, bacterial and host interactions with the aim of improving the treatment of infectious diseases. Unlike traditional microbiological tests (i.e. MICs, MBCs), in vitro pharmacodynamic models (IPDMs) characterize important relationships between antibiotic concentrations, bacterial susceptibilities and antibacterial responses. Concentration profiles in vivo of different antibiotic regimens can be studied against various isolates to produce wide ranges of PD indices, including peak concentration divided by MIC (peak/mic), time of concentration above MIC (T > MIC) and AUC 24 /MIC. PD models can be constructed to identify PD indices and thresholds associated with optimal antibacterial effect. The application of these principles to clinical trials can be used to determine... *Corresponding author. Tel: +1-204-474-8414; Fax: +1-204-474-7617; E-mail: zelenits@ms.umanitoba.ca... 905 2003 The British Society for Antimicrobial Chemotherapy

S. A. Zelenitsky et al. optimal dosing strategies for the treatment of infectious diseases. Increases in penicillin-resistant Streptococcus pneumoniae and fluoroquinolone use for respiratory tract infections have stimulated significant interest in fluoroquinolone PDs against Gram-positive bacteria. IPDMs have demonstrated the generally rapid and potent activities of simulated clinical doses of agents such as gatifloxacin, gemifloxacin, levofloxacin, moxifloxacin and trovafloxacin against susceptible S. pneumoniae isolates. 1 11 More comprehensive PD characterizations have shown concentration-dependent antibiotic activity, and further characterized relationships between peak/mic or AUC 24 /MIC and microbiological response. 12 14 AUC/MIC in vitro has been primarily investigated as an index of final or overall measures of bacterial kill. Most often, it has been associated with bacterial kill at the end of dosing intervals, or with areas related to the bacterial kill curve, such as the area between the control growth and time kill curves (ABBC), the area above the curve (AAC) and the area under the time kill curve (AUBC). Innovative work by Firsov et al. 15 found that measures of response that do not consider effect duration (i.e. time at which response is achieved) resulted in different AUC/MIC relationships. The investigators overcame this issue by using a measure of intensity of antibiotic effect (IE), or the area between control growth and time kill curves from time zero to the time of maximal regrowth, even if the time was greater than the dosing interval. 16 Our goal was to conduct a comprehensive PD analysis based on dose escalations of moxifloxacin and levofloxacin against S. pneumoniae in an IPDM. In addition to traditional PD indices (i.e. peak/mic, T > MIC, AUC 24 /MIC), we were interested in the application of the latter as a continuous index of fluoroquinolone exposure, and the potential association between AUC 0 t /MIC during the dosing interval and bacterial kill at the relative times. Materials and methods Bacterial strains Four strains of S. pneumoniae, including an ATCC (49619) isolate and three clinical (81, 63, 26) isolates from blood, were used. MICs were determined using the broth microdilution method described by the NCCLS. 17 Isolates ATCC 49619 and 63 were penicillin sensitive (MICs 0.03, 0.03 mg/l), and isolates 81 and 26 were penicillin resistant (MICs 2, 2 mg/l). Moxifloxacin and levofloxacin MICs were 0.125 and 1 mg/l for isolates ATCC 49619 and 81, and 0.25 and 0.5 mg/l for isolates 63 and 26. Isolates ATCC 49619 and 81 were used in dose escalation studies, whereas isolates 63 and 26 were used in validation experiments. Antibiotics and medium Moxifloxacin was supplied by Bayer Corporation (West Haven, CT, USA) and levofloxacin by Ortho-McNeil Pharmaceutical (Raritan, NJ, USA). Mueller Hinton broth (Difco Laboratories, Detroit, MI, USA) supplemented with calcium 25 mg/l and magnesium 12.5 mg/l, with 4% lysed horse blood (Quelab, QC, Canada) was used in all experiments (CSMHB 4%LHB). In vitro pharmacodynamic model A one-compartment IPDM was used to simulate S. pneumoniae bacteraemia in an immunocompromised host. Four central compartments consisting of 250 ml glass flasks were stirred with magnetic bars and maintained at 37 C in a heated water bath. A computerized pump (Masterflex, Cole-Parmer Instrument Company, Chicago, IL, USA) was used to program the flow of sterile CSMHB 4%LHB through the central compartments, and produce 11 and 7 h elimination half-lives for moxifloxacin and levofloxacin, respectively. 18 Flow rates were calibrated prior to experiments and were measured again at 6 and 24 h. Experiments were excluded if the variation exceeded 5% of the original flow rate. Bacterial suspensions were prepared with the inoculation of 15 20 colonies from solid blood agar into 100 ml of broth. Suspensions were incubated for 4 h, adjusted to a 0.5 McFarland s standard density and injected into the central compartments to yield initial inocula of 1 10 6 cfu/ml at time zero. Dose escalation studies were performed by injecting single bolus doses of moxifloxacin and levofloxacin, with peak concentrations equivalent to 1, 2, 4, 8, 16 and 32 MIC, against isolates ATCC 49619 and 81. Samples were collected at 0, 0.5, 1, 2, 6 and 24 h, serially diluted in normal saline at 4 C and aliquotted (10 and 100 µl) onto solid blood agar in duplicate. Plates were incubated for 24 h and then viable colonies between 10 and 100 per plate were counted. The lower limit of detection was 1 10 2 cfu/ml. To identify potential antibiotic carry-over, concurrent samples underwent two cycles of washing to remove antibiotic. Washing was performed by centrifuging at 4000g for 10 min, decanting the supernatant and suspending the remaining pellet in broth to the initial volume. Colony counts from washed and unwashed samples were compared. All experiments were performed with a growth control and in triplicate on separate occasions. Pharmacodynamic analysis Initial inocula were standardized to 1 10 6 cfu/ml, and bacterial kill curves were constructed using mean log 10 cfu/ml (for experiments performed in triplicate) versus time. Twentyfour bacterial kill curves were produced from dose escalation studies of six concentration profiles of moxifloxacin and 906

Fluoroquinolone exposure in an infection model levofloxacin against two isolates. Colony count variation at each time point within and among experiments performed in triplicate was <10%, and antibiotic carry-over effects were not observed. Bacterial kill was measured by subtracting bacterial counts at 1, 2, 6 and 24 h from initial inocula, with positive values indicating net kill and negative values representing net growth. Bactericidal activity was defined as bacterial kill >3 log 10 cfu/ml, and bacterial eradication was defined as bacterial counts below the limit of detection, or bacterial kill >4 log 10 cfu/ml. Using a one-compartment, linear pharmacokinetic model, PD indices were estimated for the six concentration profiles of moxifloxacin and levofloxacin against isolates ATCC 49619 and 81. Peak/MIC was calculated as the peak concentration divided by MIC, T > MIC was the percentage of time of concentration above MIC for the 24 h dosing interval, and AUC 24 /MIC was the area under the concentration time curve from 0 to 24 h divided by MIC. For moxifloxacin, peak/mics of 1, 2, 4, 8, 16 and 32 produced T > MICs of 0, 46%, 92%, 100%, 100% and 100%, and AUC 24 /MICs of 12, 25, 50, 99, 198 and 396, respectively. For levofloxacin, the same peak/ MICs produced T > MICs of 0, 29%, 58%, 88%, 100% and 100%, and AUC 24 /MICs of 9, 18, 37, 73,147 and 293. As a continuous index of antibiotic exposure, AUC 0 t /MIC was also calculated as AUC from 0 to 1, 2 and 6 h divided by MIC (Figure 1). AUC 0 t /MICs, including AUC 0 1 /MIC, AUC 0 2 / MIC, AUC 0 6 /MIC and AUC 0 24 /MIC, ranged from 0.97 to 396 for moxifloxacin and 0.95 to 293 for levofloxacin. Relationships between the PD indices and bacterial kill were analysed using linear and non-linear (e.g. sigmoidal E max ) methods where appropriate. Goodness of fit for non-linear models was assessed using root mean squares, residual plots and Akaike s Information Criteria. The PD model was validated using simulated clinical doses of moxifloxacin 400 mg once daily and levofloxacin 500 mg once daily against two different clinical isolates, 63 and 26. In the IPDM, unbound concentration profiles for moxifloxacin (peak 2.3 mg/l, trough 0.5 mg/l, AUC 24 29 mg h/l) and levofloxacin (peak 4.2 mg/l, trough 0.4 mg/l, AUC 24 38 mg h/l) in adults with normal renal function were simulated. 18 AUC 0 t /MIC at 1, 2, 6 and 24 h were estimated, and used to predict bacterial kill curve responses. Measures of precision (i.e. root mean square error) and bias (i.e. mean error) were used to assess predictive performance. Statistical computations were executed using SPSS for Windows, release 11 software (SPSS Inc.). Results Representative bacterial kill curves for the dose escalation studies of moxifloxacin and levofloxacin against isolate 81 are shown in Figure 2. Similar results were observed against Figure 1. AUC 0 t /MIC as a continuous index of antibiotic exposure and antibacterial response. The upper broken line is the bacterial kill curve, the continuous line is the antibiotic concentration profile and the lower broken line is the MIC. Figure 2. Representative bacterial kill curves for dose escalation studies of moxifloxacin (a) and levofloxacin (b) against S. pneumoniae isolate 81. Curves were produced by doses with peak concentrations equivalent to 1 (open squares), 2 (open triangles), 4 (open circles), 8 (filled triangles), 16 (filled squares) and 32 (filled circles) MIC. 907

S. A. Zelenitsky et al. Figure 3. PD model of peak/mic and bacterial kill (log 10 cfu/ml) at 1 h. Data represent moxifloxacin (circles) and levofloxacin (squares) against S. pneumoniae isolates ATCC 49619 and 81 (S = 1.5, K m = 9.3, r 2 = 0.95). Figure 4. PD model of T > MIC and bacterial kill (log 10 cfu/ml) at 24 h. Data represent moxifloxacin (circles) and levofloxacin (squares) against S. pneumoniae isolates ATCC 49619 and 81 (S = 20.8, K m = 51.4, r 2 = 0.99). isolate ATCC 49619. Concentration-dependent activity was observed for both antibiotics. There was little to no bacterial inhibition when initial peak concentrations were equivalent to the MIC, and limited kill followed by regrowth when peaks were 2 MIC. All other doses, however, were at least bactericidal with >3 log 10 cfu/ml of bacterial kill at 24 h. Although peak/mic was associated with bacterial kill at all times, the relationship was strongest at 1 (Figure 3), 2 and 6 h, but less descriptive at 24 h. Peak/MIC also demonstrated significant influence on the time to bactericidal effect. Based on the PD model, peak/mics of 17 are bactericidal at 2 h, peak/mics of 5 are bactericidal at 6 h and peak/mics of 3 are not bactericidal until 24 h. T > MIC showed a strong association with final bacterial kill at 24 h (Figure 4). The steep PD response curve predicts that T > MICs of 48% and 55% are bacteriostatic and bactericidal, respectively. Furthermore, T > MICs over 70% result in bacterial eradication at 24 h. AUC 0 t /MIC exhibited the strongest association with bacterial kill during (i.e. at 1, 2 and 6 h) and at the end of dosing intervals (i.e. at 24 h) until regrowth or eradication. Figure 5 shows the PD model, derived from 80 data points, which Figure 5. PD model of AUC 0 t /MIC and bacterial kill (log 10 cfu/ml) where data represent estimated AUC 0 t /MIC and measured bacterial kill at 1, 2, 6 and 24 h. Data represent moxifloxacin (circles) and levofloxacin (squares) against S. pneumoniae isolates ATCC 49619 and 81 (S = 1.8, K m = 16.5, r 2 = 0.95). Figure 6. Predicted (broken line) and observed (solid lines) bacterial kill curves produced by simulated clinical doses of moxifloxacin (a) and levofloxacin (b) against S. pneumoniae isolates 63 (squares) and 26 (circles). describes the relationship between AUC 0 t /MIC at 1, 2, 6 and 24 h and bacterial kill at the respective times (S = 1.8, K m = 16.5, r 2 = 0.95). The model predicts that AUC 0 t /MICs of 28 and 135 produce 3 and 4 log 10 cfu/ml of bacterial kill, respectively. Furthermore, AUC 0 t /MIC was highly predictive of bacterial kill curves produced by simulated clinical doses of moxifloxacin and levofloxacin against S. pneumoniae isolates 63 and 26 (root mean square error 0.36 log 10 cfu/ml, mean error 0.02 log 10 cfu/ml) (Figure 6). 908

Fluoroquinolone exposure in an infection model Discussion As the use of fluoroquinolones for respiratory tract infections increases, the application of PD principles to maximize efficacy and minimize resistance becomes more important. This study provided a comprehensive characterization of traditional (i.e. peak/mic, T > MIC, AUC 24 /MIC) and novel (i.e. AUC 0 t /MIC) PD indices. In vitro, 12,13 animal 19,20 and human 21 studies of fluoroquinolones against S. pneumoniae have established AUC/MIC as a reliable index of response. AUC/MIC in vitro has been primarily investigated as an index of final or overall measures of bacterial kill. Traditionally, AUC 24 /MIC has been associated with bacterial kill at 24 h, or areas related to the bacterial kill curves over the dosing intervals (e.g. ABBC, AAC, AUBC). In this study, AUC 0 t /MIC was used as a continuous index of fluoroquinolone exposure, which predicted bacterial kill throughout the dosing interval until eradication or regrowth. Since AUC 0 t /MIC incorporated time as an independent variable, it predicted not only the effect, but also the time at which the response was achieved. The PD model predicts bactericidal activity when AUC 0 t /MICs are 28, and bacterial eradication when values exceed 135. It could be that, based on different pharmacokinetic and potency profiles, fluoroquinolone A achieves an AUC/MIC of 28 at 2 h, whereas agent B reaches this value at 12 h. The PD model would then predict a bacterial kill of 3 log 10 cfu/ml at 2 h with agent A, but not until 12 h with agent B. Furthermore, if agent A achieves an AUC/MIC of 135 at 12 h and agent B fails to reach this value during the dosing interval, the model would predict bacterial eradication with agent A only at 12 h. The lower threshold (i.e. 28) in this study is similar to that from previous reports. Optimal antibacterial activity has been associated with AUC 24 /MICs > 30 50, 2 4,7,12 whereas suboptimal effects with resistance or regrowth have been observed with AUC 24 /MICs < 20 30. 8,12,14 There is some variability in thresholds for bacterial eradication, which may be due to study isolates, ranges of AUC/MICs or definitions of response. Whereas some reports found bacterial eradication with AUC 24 /MICs > 30, 10,12 other studies, such as this one, demonstrated higher thresholds for complete antibacterial responses. 13,14 In a study of simulated clinical doses of six fluoroquinolones against two laboratory-derived ciprofloxacin-resistant isolates of S. pneumoniae, Coyle et al. 14 determined that bacterial counts below detection at 24 h were associated with AUC 24 /MICs > 82. Another investigation, which tested three doses of gemifloxacin against five isolates of S. pneumoniae, found maximal AUBCs at 48 h when AUC/ MICs were between 300 and 400. 13 As previously discussed, AUC 0 t /MIC expands the application of these thresholds from associations with final or overall antibacterial response to a PD index, which predicts both the magnitude and time of a desired effect. In other words, AUC 0 t /MIC considers potential differences between agents that may both achieve an AUC/MIC of 30, but where one reaches it at 2 h and the other at 24 h. The clinical application of AUC 24 /MIC has been demonstrated, whereas the relevance of AUC 0 t /MIC remains to be seen. Although optimal fluoroquinolone AUC 24 /MICs of 125 250 are suggested for Gram-negative pathogens, 22 lower thresholds have been found for S. pneumoniae infections. One study of pneumococcal pneumonia in mice found a significant AUC 24 /MIC threshold of >160, 23 whereas a second animal study reported optimal responses, with values of 30 40. 20 Ambrose et al. 21 also found a lower threshold based on the analysis of data from two Phase III trials of levofloxacin versus gatifloxacin for community-acquired pneumonia or acute bronchitis involving S. pneumoniae. The probability of microbiological response was 100% compared with 64% when unbound AUC 24 /MICs were above and below 33.7, respectively. Although lower thresholds appear effective, there are concerns that such targets may promote bacterial resistance. The characterization of antibiotic PDs is often complicated by the strong co-variance among indices. 18,24 In this study, peak/mic correlated well with early bacterial kill, but was less indicative of later bacterial counts at 24 h. Peak/MIC is founded on the concentration-dependent activity of fluoroquinolones, but does not consider antibiotic elimination and becomes less predictive of antibacterial response over time. On the other hand, AUC 0 t /MIC includes components of time and antibiotic elimination, and thereby maintains its predictive performance throughout the dosing interval. Finally, T > MIC is associated with fluoroquinolone activity, especially when concentrations fall below the MIC. 25 In contrast to peak/mic, T > MIC is most predictive of later bacterial counts at the end of dosing intervals. In this study, maximal antibacterial response at 24 h was associated with T > MICs > 70% (17 in 24 h). Although this threshold is similar to those of other IPDM studies, 7,13 it differs from another, in which bacterial eradication was observed when T > MICs exceeded 8 in 24 h (33%). 12 As mentioned previously, such variability may be explained by study isolates, ranges of T > MICs or definitions of response. This study has some important limitations. First, the PD characterizations were based on fluoroquinolone activity against two sensitive S. pneumoniae isolates. Dose escalations produced a wide range of values for most PD indices; however, data were less comprehensive for parameters such as T > MIC. In addition, AUC 0 t /MIC and the PD model would benefit from further validation using isolates with MICs outside the relatively narrow range of those used in this study. Although moxifloxacin and levofloxacin demonstrated the same PD responses, the study of other fluoroquinolones would also broaden the application. The lack of sample times between 6 and 24 h would not have added significantly to the 909

S. A. Zelenitsky et al. bacterial kill curves, but would have allowed validation at other times. Finally, the application of this PD model, in the setting of multiple dose regimens, requires further investigation. In conclusion, this study demonstrated the novel application of AUC 0 t /MIC as a continuous index of fluoroquinolone activity. It was predictive of bacterial kill until eradication or regrowth, and provided an extensive characterization of antibiotic PDs against S. pneumoniae. Acknowledgements This project was supported in part by Bayer Corporation. References 1. Bowker, K. E., Wootton, M., Rogers, C. A., Lewis, R., Holt, H. A. & MacGowan, A. P. (1999). Comparison of in-vitro pharmacodynamics of once and twice daily ciprofloxacin. Journal of Antimicrobial Chemotherapy 44, 661 7. 2. Lacy, M. K., Lu, W., Xu, X., Tessier, P. R., Nicolau, D. P., Quintiliani, R. et al. (1999). Pharmacodynamic comparisons of levofloxacin, ciprofloxacin, and ampicillin against Streptococcus pneumoniae in an in vitro model of infection. Antimicrobial Agents and Chemotherapy 43, 672 7. 3. Lister, P. D. & Sanders, C. C. (1999). Pharmacodynamics of trovafloxacin, ofloxacin, and ciprofloxacin against Streptococcus pneumoniae in an in vitro pharmacokinetic model. Antimicrobial Agents and Chemotherapy 43, 1118 23. 4. Lister, P. D. & Sanders, C. C. (1999). Pharmacodynamics of levofloxacin and ciprofloxacin against Streptococcus pneumoniae. Journal of Antimicrobial Chemotherapy 43, 79 86. 5. MacGowan, A. P., Bowker, K. E., Wootton, M. & Holt, H. A. (1999). Activity of moxifloxacin, administered once a day, against Streptococcus pneumoniae in an in vitro pharmacodynamic model of infection. Antimicrobial Agents and Chemotherapy 43, 1560 4. 6. Esposito, S., Noviello, S. & Ianniello, F. (2000). Bactericidal activity of moxifloxacin compared to grepafloxacin and clarithromycin against Streptococcus pneumoniae and Streptococcus pyogenes investigated using an in vitro pharmacodynamic model. Journal of Chemotherapy 12, 475 81. 7. Hershberger, E. & Rybak, M. J. (2000). Activities of trovafloxacin, gatifloxacin, clinafloxacin, sparfloxacin, levofloxacin, and ciprofloxacin against penicillin-resistant Streptococcus pneumoniae in an in vitro infection model. Antimicrobial Agents and Chemotherapy 44, 598 601. 8. Madaras-Kelly, K. J. & Demasters, T. A. (2000). In vitro characterization of fluoroquinolone concentration/mic antimicrobial activity and resistance while simulating clinical pharmacokinetics of levofloxacin, ofloxacin, or ciprofloxacin against Streptococcus pneumoniae. Diagnostic Microbiology and Infectious Disease 37, 253 60. 9. Zinner, S. H., Simmons, K. & Gilbert, D. (2000). Comparative activities of ciprofloxacin and levofloxacin against Streptococcus pneumoniae in an in vitro dynamic model. Antimicrobial Agents and Chemotherapy 44, 773 4. 10. Lister, P. D. & Sanders, C. C. (2001). Pharmacodynamics of moxifloxacin, levofloxacin and sparfloxacin against Streptococcus pneumoniae. Journal of Antimicrobial Chemotherapy 47, 811 8. 11. Zhanel, G. G., Walters, M., Laing, N. & Hoban, D. J. (2001). In vitro pharmacodynamic modelling simulating free serum concentrations of fluoroquinolones against multidrug-resistant Streptococcus pneumoniae. Journal of Antimicrobial Chemotherapy 47, 435 40. 12. Lister, P. D. (2002). Pharmacodynamics of gatifloxacin against Streptococcus pneumoniae in an in vitro pharmacokinetic model: impact of area under the curve/mic ratios on eradication. Antimicrobial Agents and Chemotherapy 46, 69 74. 13. MacGowan, A. P., Rogers, C. A., Holt, H. A., Wootton, M. & Bowker, K. E. (2001). Pharmacodynamics of gemifloxacin against Streptococcus pneumoniae in an in vitro pharmacokinetic model of infection. Antimicrobial Agents and Chemotherapy 45, 2916 21. 14. Coyle, E. A., Kaatz, G. W. & Rybak, M. J. (2001). Activities of newer fluoroquinolones against ciprofloxacin-resistant Streptococcus pneumoniae. Antimicrobial Agents and Chemotherapy 45, 1654 9. 15. Firsov, A. A., Lubenko, I. Y., Portnoy, Y. A., Zinner, S. H. & Vostrov, S. N. (2001). Relationships of the area under the curve/ MIC ratio to different integral endpoints of the antimicrobial effect: gemifloxacin pharmacodynamics in an in vitro dynamic model. Antimicrobial Agents and Chemotherapy 45, 927 31. 16. Firsov, A. A., Zinner, S. H., Vostrov, S. N., Portnoy, Y. A. & Lubenko, I. Y. (2002). AUC/MIC relationships to different endpoints of the antimicrobial effect: multiple-dose in vitro simulations with moxifloxacin and levofloxacin. Journal of Antimicrobial Chemotherapy 50, 533 9. 17. National Committee for Clinical Laboratory Standards. (2000). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically Fifth Edition: Approved Standard M7-A5. NCCLS, Wayne, PA, USA. 18. Wright, D. H., Brown, G. H., Peterson, M. L. & Rotschafer, J. C. (2000). Application of fluoroquinolone pharmacodynamics. Journal of Antimicrobial Chemotherapy 46, 669 83. 19. Ng, W., Lutsar, I., Wubbel, L., Ghaffar, F., Jafri, H., McCracken, G. H. et al. (1999). Pharmacodynamics of trovafloxacin in a mouse model of cephalosporin-resistant Streptococcus pneumoniae pneumonia. Journal of Antimicrobial Chemotherapy 43, 811 6. 20. Mattoes, H. M., Banevicius, M., Li, D., Turley, C., Xuan, D., Nightingale, C. H. et al. (2001). Pharmacodynamic assessment of gatifloxacin against Streptococcus pneumoniae. Antimicrobial Agents and Chemotherapy 45, 2092 7. 21. Ambrose, P. G., Grasela, D. M., Grasela, T. H., Passarell, J., Mayer, H. B. & Pierce, P. F. (2001). Pharmacodynamics of fluoroquinolones against Streptococcus pneumoniae in patients with community-acquired respiratory tract infections. Antimicrobial Agents and Chemotherapy 45, 2793 7. 22. Forrest, A., Nix, D. E., Ballow, C. H., Goss, T. F., Birmingham, M. C. & Schentag, J. J. (1993). Pharmacodynamics of intravenous ciprofloxacin in seriously ill patients. Antimicrobial Agents and Chemotherapy 37, 1073 81. 910

Fluoroquinolone exposure in an infection model 23. Bedos, J. P., Azoulay-Dupuis, E., Moine, P., Muffat-Joly, M., Veber, B., Pocidalo, J. J. et al. (1998). Pharmacodynamic activities of ciprofloxacin and sparfloxacin in a murine pneumococcal pneumonia model: relevance for drug efficacy. Journal of Pharmacology and Experimental Therapeutics 286, 29 35. 24. MacGowan, A., Rogers, C. & Bowker, K. (2000). The use of in vitro pharmacodynamic models of infection to optimize fluoroquinolone dosing regimens. Journal of Antimicrobial Chemotherapy 46, 163 70. 25. Sanchez-Recio, M. M., Colino, C. I. & Sanchez-Navarro, A. (2000). A retrospective analysis of pharmacokinetic/pharmacodynamic indices as indicators of the clinical efficacy of ciprofloxacin. Journal of Antimicrobial Chemotherapy 45, 321 8. 911