JAC Bactericidal index: a new way to assess quinolone bactericidal activity in vitro

Journal of Antimicrobial Chemotherapy (1997) 39, 713 717 JAC Bactericidal index: a new way to assess quinolone bactericidal activity in vitro Ian Morrissey* Department of Biosciences, Division of Biochemistry & Microbiology, University of Hertfordshire, Hatfield, Herts AL10 9AB, UK A new method of assessing the bactericidal activity of quinolones and other antibacterial agents, known as the bactericidal index or BI, is introduced and discussed. The BI represents total bacterial kill over a drug concentration range limited to levels achievable in vivo. This method allows the bactericidal activity of drugs to be readily compared. Calculation of BI produces a single value per bacterial strain/antibacterial agent combination that represents overall bactericidal activity at clinically relevant concentrations. As an example, the bactericidal activities of quinolones against Enterococcus faecalis are compared. Introduction Since the introduction of nalidixic acid in the 1960s, many thousands of quinolone derivatives have been synthesized and analysed for antibacterial activity. A few of these drugs have been developed and marketed for use. All these quinolones are highly bactericidal against susceptible strains. 1 As with all antimicrobial agents, initial investigations begin with MIC determinations. This convenient and established procedure allows comparisons to be made with other antibacterial agents. Unfortunately, MICs only measure inhibition of bacterial multiplication and do not reflect bactericidal activity. With bacteriostatic antibacterial agents such as chloramphenicol, MIC determinations may be sufficient. With highly bactericidal drugs such as the quinolones it is unwise to rely on MIC alone. For example, the MICs of quinolones against Pseudomonas aeruginosa are relatively high compared with MICs against other Gram-negative bacteria. If these MIC data are extended to predict bactericidal activity, the quinolones would be expected to show poor kill against P. aeruginosa. In fact, quite the opposite is true. Ciprofloxacin, ofloxacin and levofloxacin are highly bactericidal against P. aerugi - nosa despite having relatively high MICs. 2 In addition, there are distinct differences between quinolone activity at the MIC and activity at a bactericidal concentration. With ciprofloxacin, acid ph increases MIC significantly, but bactericidal activity is less affected. Conversely, magnesium ions antagonize the bacteriostatic effects of quinolones much less than their bactericidal action. 3 This effect may occur because the quinolones show additional bactericidal mechanisms of action at concentrations higher than the MIC. 4 Therefore, it is important to distinguish between the bacteriostatic and bactericidal activities of the quinolone antibacterial agents. The bactericidal activity of quinolones has been well documented, with many of these drugs being assessed by the method introduced by Professor J. T. Smith. 1 The technique involves calculating the per cent survival of bacteria treated over a fixed range of quinolone concentrations for 3 h. A biphasic dose response occurs with most quinolones, producing a single concentration of maximum kill. This concentration is known as the optimum bactericidal concentration (OBC) and OBCs have been used to compare the bactericidal activity of quinolones. Less potent older quinolones such as nalidixic acid have considerably higher OBCs compared with fluoroquinolones such as ciprofloxacin. Hence, it is clear which quinolones are the more potent by simply comparing OBCs. Unfortunately, more recently developed quinolones tend to have very similar OBCs, which makes comparisons between these drugs difficult. In addition, other classes of antibacterial agent do not produce OBCs at all thereby making it impossible to compare quinolones with nonquinolones using OBC data. OBC comparison alone totally disregards the extent of bactericidal activity that occurs at clinically achievable quinolone concentrations higher or lower than the OBC. Drugs such as nalidixic acid have OBCs much greater than concentrations that can be *Tel: 44-1707-285163; Fax: 44-1707-2885046; E-mail: i.morrissey herts.ac.uk 1997 The British Society for Antimicrobial Chemotherapy 713

I. Morrissey achieved in the serum and body tissues and most fluoroquinolones have OBCs against Gram-positive bacteria that are only just achievable in the serum. Thus, the use of OBC data may not be an ideal way to analyse bactericidal activity. An alternative and commonly used method of bactericidal activity assessment is that of time kill studies, which are often carried out using set multiples of the drug s MIC. This method may not be ideal for quinolones, where MIC values are unrelated to bactericidal activity as discussed above. Aesthetically, a problem with this method is that graphs of generated data need to be compared, which is not conducive to quick and easy inter-drug comparisons. This becomes especially difficult when the drugs in question show large differences in their MICs and therefore would be tested and compared for bactericidal activity using unsuitable concentrations. Another common assessment for bactericidal activity is that of minimum bactericidal concentration (MBC), which is normally taken as the lowest drug concentration that produces 99.9% kill. This method has the advantage over time kill studies that inter-drug comparisons can be readily made. However, MBCs only indicate the initiation of bactericidal activity and not activity at higher drug concentrations. Therefore MBC measurements have similar drawbacks to OBC measurements. Therefore, there is a need for an alternative way to measure bactericidal activity to take into account all concentrations that are available during drug therapy and to produce data which allow easy comparisons between quinolones and other antibacterial agents to be made. In this paper, the bactericidal activity of the quinolones ciprofloxacin, ofloxacin, temafloxacin, DU-6859a and trovafloxacin are compared against Enterococcus faecalis ATCC 19433 by calculating bacterial indexes (BIs). E. faecalis was chosen for this study because of its inherent resistance to quinolones, producing slow quinolone bactericidal activity 5 and also because E. faecalis does not always show a distinct OBC. Materials and methods Determination of bactericidal activity (bactericidal profiles) Quinolone bactericidal activity was tested against E. fae - calis ATCC 19433 by the method of Morrissey & Smith. 6 Briefly, overnight cultures of E. faecalis were diluted 1 in 50 in nutrient broth no. 2 (Unipath, Basingstoke, UK) to give an initial inoculum size of about 10 7 cfu/ml. The broth was set up containing a range of quinolone concentrations between 0.015 and 50 mg/l and incubated for 24 h at 37 C. Viable counts were taken at 3, 6 and 24 h. In contrast to the method of Morrissey & Smith, 6 results were expressed as viable count (cfu/ml) rather than per cent survival. Calculation of BIs Results obtained from the bactericidal profiles were used to produce a BI for each quinolone against E. faecalis at 3, 6 and 24 h. The data used for DU-6859a and trovafloxacin have been published previously, 6,7 but not analysed in this way. BIs were calculated by plotting the logarithm of the reduction in cfu against the logarithm of the concentration of antimicrobial agent using data from the bactericidal profiles. The highest drug concentration considered in this plot was the peak serum concentration (C max ) taken from published literature (Table). Reduction in viability at the C max was interpolated from the bactericidal profiles. The BI was measured as the area under the curve (AUC) for the bactericidal portion of this plot. Table. C max values used to calculate bactericidal indices in this study Single oral dose C max Source of data Drug (mg) (mg/l) (reference no.) Ofloxacin 400 5.85 10 Ciprofloxacin 500 2.60 11 Trovafloxacin 300 3.00 12 DU-6859a 200 1.86 13 Temafloxacin 400 2.43 14 Results The bactericidal profiles for ciprofloxacin and DU-6859a against E. faecalis are shown in Figure 1. It can be seen that the bacteria were able to multiply at ciprofloxacin concentrations below 1.0 mg/l and at DU-6859a concentrations below 0.1 mg/l. Therefore, DU-6859a was ten times more inhibitory than ciprofloxacin. At higher concentrations, ciprofloxacin was bactericidal against E. faecalis after 6 and 24 h, but not after 3 h. DU-6859a, on the other hand, was bactericidal after just 3 h. The other quinolones tested showed bactericidal activity at a level between that of ciprofloxacin and DU-6859a. As with ciprofloxacin, all the quinolones tested with the exception of DU-6859a required 6 or 24 h incubation for bactericidal activity (results not shown). The plots of log reduction in viability against log drug concentration for ciprofloxacin and DU-6859a after 3 h are shown in Figure 2. The log reduction values for C max were interpolated from Figure 1, as these values were not tested directly. Data for 6 and 24 h have been omitted for clarity. It can be seen that DU-6859a has a considerably greater bactericidal area than ciprofloxacin. The BIs (i.e. AUC values from plots made as shown in Figure 2) for all the quinolones against E. faecalis over 3, 6 714

Quinolone bactericidal index 10 7 10 7 Figure 2. Plot of the logarithm of reduction in cfu against the logarithm of drug concentration for (a) ciprofloxacin and (b) DU-6859a against E. faecalis ATCC 19433 treated for 3 h at 37 C. against E. faecalis compared with trovafloxacin, ofloxacin, temafloxacin and ciprofloxacin. Discussion Figure 1. Bactericidal profiles for (a) ciprofloxacin and (b) DU- 6859a against E. faecalis ATCC 19433 for 3 ( ), 6 ( ) and 24 h ( ) at 37 C. and 24 h are summarized in Figure 3. It can be seen that the rank order of quinolone potency, as measured by BI, was essentially the same if bactericidal activity was measured at 3, 6 or 24 h. All the quinolones became more bactericidal with increasing incubation time. The BIs shown in Figure 3 clearly illustrate the enhanced potency of DU-6859a In this study, BI values indicated that DU-6859a has higher bactericidal activity than trovafloxacin and established quinolones such as ofloxacin and ciprofloxacin against E. faecalis. In addition, DU-6859a was unique in that it showed potent bactericidal activity after just 3 h. This ability has been noted previously and has been suggested to be due to the possession of an additional bactericidal mechanism of action by DU-6859a against E. faecalis. 6 Trovafloxacin was the next most potent quinolone as measured by BI. The improved BI compared with temafloxacin, ciprofloxacin or ofloxacin was due to the fact that trovafloxacin was bactericidal over a wider range of clinically relevant concentrations. This illustrates the usefulness of the BI method because it takes into account 715

I. Morrissey Figure 3. Histogram illustrating the bactericidal index for quinolones against Enterococcus faecalis ATCC 19433. Cip, ciprofloxacin; Tem, temafloxacin; Ofl, ofloxacin; Tro, trovafloxacin; DU, DU-6859a. bactericidal potential over a wide range of concentrations. It is interesting to note that the OBC for DU-6859a against E. faecalis is 1.5 mg/l, 6 but for trovafloxacin the OBC is 0.9 mg/l. 7 Using OBC alone, therefore, one might think that trovafloxacin is superior to DU-6859a. BI data, however, show that trovafloxacin is not as bactericidal as DU-6859a when a wider range of concentrations is taken into account. BIs were calculated by estimating viable count using a spread-plating technique followed by calculation of reduction in viability due to drug treatment. Other methods of measuring viability may also be employed to calculate BIs, such as the micro-method suggested recently. 8 However, viability assessment by measuring optical density would not be appropriate owing to filamentation of bacteria upon quinolone treatment. 1 As the BI relies on calculation of an AUC, the drug concentrations used need not be those in this study. The only requirement is that the highest drug concentration tested must exceed the C max for each test drug. To avoid the need to interpolate, the C max may be included as a test concentration. The main purpose of this paper was to introduce a new method of measuring bactericidal potency that overcomes the difficulties associated with other bactericidal methods. A method was required that produces a single indicator to compare the bactericidal activity of one drug with another, but at the same time reflect bacterial killing at clinically achievable concentrations. The measurement and calculation of BI achieves both these aims, where the higher the BI the more bactericidal the drug should be in a clinical setting. This BI method has been successfully applied to compare the bactericidal activity of three quinolones and cefotaxime against Streptococcus pneumoniae. 9 Nevertheless, BI results need to be compared with future clinical results to verify this new method of assessing antimicrobial agents. In this study, BIs were evaluated using C max values from single oral doses as the cut-off point in calculating total bactericidal activity. This method can be very easily adapted to assess BIs with pharmacokinetic data from multiple dosing regimens, where higher drug levels can be reached. Due to the biphasic nature of dose response curves to quinolones, 4 higher dose levels will not increase the bactericidal potency of quinolones uniformly, but will have a consistent effect on the BI. Additionally, the BI method can be used to assess bactericidal activity at concentrations expected at particular sites of infection, such as the urinary tract or CSF. It may be possible to predict the bactericidal action of drugs against specific infections in patients who have altered drug pharmacokinetics by calculating BIs with suitably altered cut-off points. Furthermore, the potential bactericidal effects of alternative drug therapies may be assessed for drug-resistant bacterial strains directly isolated from patients. Acknowledgement This paper is dedicated to the memory of Professor J. T. Smith, who died on 15th July 1996. His death is a great loss to the world of antimicrobial chemotherapy. 716

Quinolone bactericidal index References 1. Smith, J. T. (1984). Awakening the slumbering potential of the 4-quinolone antibacterials. Pharmaceutical Journal 233, 299 305. 2. Morrissey, I. & Smith, J. T. (1994). Activity of 4-quinolones against Pseudomonas aeruginosa. Arzneimittel-Forschung Drug Research 44, 1157 61. 3. Smith, J. T. & Lewin, C. S. (1988). Chemistry and mechanisms of action of the quinolone antibacterials. In The Quinolones (Andriole, V. T., Ed.), pp. 23 82. Academic Press, London. 4. Lewin, C. S., Morrissey, I. & Smith, J. T. (1991). The mode of action of quinolones: the paradox in activity of low and high concentrations and activity in the anaerobic environment. European Journal of Clinical Microbiology & Infectious Diseases, 10, 240 8. 5. Lewin, C. S., Morrissey, I. & Smith, J. T. (1991). The fluoroquinolones exert a reduced rate of kill against Enterococcus faecalis. Journal of Pharmacy and Pharmacology 43, 492 4. 6. Morrissey, I. & Smith, J. T. (1995). Bactericidal activity of the new 4-quinolones DU-6859a and DV-7751a. Journal of Medical Microbiology 43, 4 8. 7. Morrissey, I. (1996). Bactericidal activity of trovafloxacin (CP-99,219). Journal of Antimicrobial Chemotherapy, 38, 1061 6. 8. Vedel, G., Bouchet, E., Gangneux, J. P. & Névot, P. (1986). A simple micro-method for time kill studies amenable to routine laboratory use. Journal of Antimicrobial Chemotherapy 37, 842 4. 9. George, J. T. & Morrissey, I. (1997). The bactericidal activity of levofloxacin compared with ofloxacin, D-ofloxacin, ciprofloxacin, sparfloxacin and cefotaxime against Streptococcus pneumoniae. Journal of Antimicrobial Chemotherapy 39, 719 23. 10. Zhang, Y., Zhang, Q., Mu, Y., Shi, Y., Wu, P. & Wang, F. (1991). Pharmacokinetics of ofloxacin in volunteers after oral administration of various single and multiple doses. In Proceedings of the 3rd International Symposium on New Quinolones. European Journal of Clinical Microbiology & Infectious Diseases, Special Issue, 254 5. 11. Gonzalez, M. A., Uribe, F., Moisen, S. D., Fuster, A. P., Selen, A., Welling, P. G. et al. (1984). Multiple-dose pharmacokinetics and safety of ciprofloxacin in normal volunteers. Antimicrobial Agents and Chemotherapy 26, 741 4. 12. Teng, R., Liston, T. E. & Harris, S. C. (1996). Multiple-dose pharmacokinetics and safety of trovafloxacin in healthy volunteers. Journal of Antimicrobial Chemotherapy 37, 955 63. 13. Nakashima, M., Uematsu, T., Kosuge, K., Umemura, K., Hakusui, H. & Tanaka, M. (1995). Pharmacokinetics and tolerance of DU-6859a, a new fluoroquinolone, after single and multiple oral doses in healthy volunteers. Antimicrobial Agents and Chemo - therapy 39, 170 4. 14. Granneman, G. R., Carpentier, P., Morrison, P. J. & Pernet, A. G. (1991). Pharmacokinetics of temafloxacin in humans after single oral doses. Antimicrobial Agents and Chemotherapy 35, 436 41. Received 12 August 1996; returned 26 September 1996; revised 4 November 1996; accepted 17 December 1996 717