- Email: [email protected]
New pharmacodynamic parameters for antimicrobial agents
International Journal of Antimicrobial Agents 13 (2000) 229 – 235
www.ischemo.org
Review
New pharmacodynamic parameters for antimicrobial agents Ronald C. Li * Pharmacokinetic/Pharmacodynamic Sciences, Genetics Institute, One Burtt Road, Ando6er, MA 01810, USA Received 11 May 1999; accepted 14 May 1999
Abstract The application of pharmacodynamic theories to antimicrobial chemotherapy has greatly improved the prediction of the time course of activity expressed by antibiotics. Being a major component of the antibiotic – bacterium interaction system, pharmacodynamics, when properly integrated with the pharmacokinetics established for the antibiotic, allow better evaluation of the dosage regimen in conjunction with its clinical response. Before this approach becomes effective, detailed background information on the complex antibiotic–bacterium interactions have to be secured. To achieve this, proper characterization of a time – kill curve is a prerequisite. The use of susceptibility endpoints such as the MIC with respect to the antibiotic concentrations achievable in vivo represent the conventional approach to clinical dosing of antimicrobial agents, i.e. by maintaining concentrations above the MIC. Recently, a number of surrogate markers have been proposed by combining suitable pharmacokinetic parameters and susceptibility data, e.g. peak/MIC ratio, AUC \ MIC, time above MIC, AUIC etc. to enhance the prediction of clinical outcomes. Attempts have been made to apply these pharmacokinetic/pharmacodynamic markers to antibiotics of the same class as well as to antibiotics from different classes. This review aims to discuss the various microbial dynamic responses in relation to antibiotic exposure and the development of different pharmacokinetic/pharmacodynamic markers for use in current antimicrobial chemotherapy. © 2000 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. Keywords: Antibiotic – bacterium interaction system; Time–kill curve; Pharmacokinetic factors
1. Introduction Over the past 2 to 3 decades, there has been a steady yet persistent increase in the use of pharmacodynamics to describe the time course of antimicrobial activity exhibited by various antibiotics. Part of the reason is the need to improve accuracy for delineating the complex interactions that exist between the antibiotic and the pathogen. As antibiotics are not instant poisons, the sequence of events demonstrating the action(s) of the antibiotic and the reaction(s) of the microorganisms usually extends over a period of time. To appreciate fully the complexity of the interactions, understanding the fundamental elements in relation to time is therefore vital not only from the perspective of evaluating the activity profiles of currently available antibiotics * Corresponding author. Tel.: +1-978-2471884; fax: + 1-9782471389. E-mail address: [email protected] (R.C. Li)
but the design of newer ones. In addition, it has always been the intent of researchers as well as clinicians to derive and extrapolate pharmacodynamic parameters in order to predict therapeutic responses more accurately.
2. Routine susceptibility testing The minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) are conventional measures to aid proper antibiotic selection for clinical use. The MIC measurement involves subjective determination of turbidity after the isolated pathogen is exposed to graded two-fold dilutions of the antibiotic for 18–24 h at 35–37°C in broth [1] whereas, MBC is the lowest concentration that produces a \99.99% reduction in viable bacterial counts in comparison to that prior to antibiotic exposure. It is apparent that these two measurements provide a single time point
0924-8579/00/$20 © 2000 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. PII: S 0 9 2 4 - 8 5 7 9 ( 9 9 ) 0 0 1 1 4 - 4
230
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
observation with no information on the temporal changes in viable bacterial cell density over the incubation period. As expected, functional use of these measurements in defining microbial pharmacodynamics is minimal.
3. Time–kill curves: pharmacodynamics and basic pharmacology The diversity of various bacterial responses to antibiotics cannot be easily understood without the establishment of a time–kill curve. A typical time – kill curve generated by subjecting the test organism to exposure at constant antibiotic concentrations, normally consists of a lag phase, a log-linear killing phase a second lag phase and a regrowth phase [2 – 5]. A static phase can be observed when resistant bacterial subpopulations overgrow to a saturated level when nutrients in the medium are exhausted. Understanding the sequence of events that takes place between the antibiotic and bacteria is vital in order to interpret the dynamics of these individual phases. Entry of the antibiotic into bacterial cells is an important first step for subsequent expression of antimicrobial activity. The cell wall/membrane usually offers an initial defensive mechanism against antibiotic permeation. For a
Fig. 1. Impact of the decreasing rate of resistance formation on the time–kill curve. When the rate of bactericidal effect remained identical at 5 h − 1 for all simulations, the reduction in the rate of resistance formation (values shown in graph) caused a delay in the development and overgrowth of resistant subpopulations. Data adapted from [2].
rapidly bactericidal antibiotic, permeation is probably a rate-limiting step for the expression of cell death. For less bactericidal antibiotics, the time lag observed between initiation of antibiotic exposure and commencement of bacterial killing may represent, in addition to permeation, the time required for all the intermediate steps that take place between receptor binding and cell death [2,3,5]. Regardless of the exact limiting factor, a high antibiotic concentration is an overriding force for shortening this lag by producing a concentration gradient favourable for cellular permeation as well as equilibration towards higher receptor occupancy. From a pharmacological perspective, the log-linear killing phase has been considered to be the most relevant indicator for bactericidal activity, i.e. as a quantitative measure for the intensity of bactericidal effect. Due to the fact that bacterial killing is the ultimate consequence following binding of the antibiotic molecules to their intended receptors and in line with most pharmacological effects, the bactericidal rate constants are saturable at high concentrations [2,3]. Generation of the concentration-response relationships between different antibiotics sharing the same mechanism of action allows direct comparison of their intrinsic potencies [2,3,5–8]. Similarly, bactericidal rate constants also permit the decrease of viable cell counts in the killing phase to be more precisely predicted during antibiotic exposure. In comparison, the regrowth phase has not been studied as extensively. Although bacterial regrowth is not a desirable effect during antibiotic exposure, it does reflect the ability of the organism to adapt to the antibiotic assault. For a number of b-lactams, the rate of bacterial resistance formation has been shown to reduce exponentially at increasing antibiotic concentrations [2]. In other words, the probability for resistant subpopulations to develop and overgrow is drastically reduced at higher antibiotic concentrations. Antibiotics showing a low potential of resistance formation should allow the bactericidal phase to be extended over a longer time for more extensive cell kill (Fig. 1). Therefore, the establishment of a full pharmacodynamic profile over time can be very informative for evaluation of the different aspects of antibiotic activity and the dynamics associated with the antibiotic-bacterium interaction. In principle, proper interpretation of various parameters describing the time–kill curve should also effectively enhance the design of new antibiotics in terms of cell permeation and receptor binding as well as resistance formation.
4. Beyond time–kill curve: postantibiotic antibiotic effect Characterization of a time–kill curve has not satisfied the quest for more endpoints describing other
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
231
Fig. 2. Antibiotic-bacterium-host interactions that exist between administration of a dose of antibiotic and the expression of antimicrobial activity.
important components of the microbial dynamic system. The persistent suppression of the bacterial growth which follows antibiotic exposure is now known as the postantibiotic effect (PAE) and is one of the more relevant pharmacodynamic parameters with respect to the optimization of dosage regimen. The PAE is highly dependent on the antibiotic – bacterial combination under investigation [9,10]. Theoretically, PAE offers a good argument for extending the dosing interval for any antibiotic exhibiting such effect. This interpretation is based on the fact that growth suppression remains when antibiotic concentration fell below its respective MIC value. Indeed, clinical utilization of PAE has been shown to be successful in aminoglycoside therapy by replacing the traditional divided daily dosing with the high dose once daily dosing regimen [11,12]. Other phenomena relating to PAE which include growth suppression following antibiotic exposure at subinhibitory concentrations (PAESME) and the enhancement of phagocytic activity during the PAE phase, also point to the functionality of PAE in designing dosage regimen for antibiotics [13,14]. The inclusion of PAE in the design of a dosing regimen cannot be completed without the consideration of pharmacokinetics as an integral factor. A recent study has demonstrated that the PAE induced is relatively longer for organisms exposed to constant antibiotic concentrations rather than exponentially decreasing concentrations even at a similar degree of antibiotic exposure in terms of AUC \ MIC [15]. This observation suggests the expression of PAE is dependent not only on the magnitude but also the mode of antibiotic exposure. Data collected from a subsequent study investigating the length of dosing interval as an independent variable suggest the state of cellular damage/recovery is a relevant determinant for the PAE observed [16,17]. These previous observations on PAE clearly demonstrated the contribution of pharmacokinetics to microbial pharmacodynamics. Unequivocally, the impact of pharmacokinetics on pharmacodynamics should be clearly defined in order to permit more accurate extrapolation of in vitro data. However, for most pharmacodynamic studies, the constraint of em-
ploying a static environment of constant antibiotic concentrations remains.
5. Pharmacokinetics: a determinant for pharmacodynamics in antimicrobial chemotherapy Pharmacokinetics are tools to depict the fundamental processes of absorption, distribution, metabolism and elimination (ADME) of any xenobiotics administered to a biological system. From the perspective of antimicrobial therapy, these processes in turn dictate the concentration versus time profile of an antibiotic and its activity in vivo. For any antibiotic to express its activity, absorption is a prerequisite to achieve adequate systemic presence in relation to the level of susceptibility. After confirming an adequate degree of absorption, the most basic approach in early day antimicrobial chemotherapy has been simply to maintain systemic levels above the MIC by varying the dose and/or dosing regimen. As for the treatment of deep tissue infections, the ability for the antibiotic to distribute or penetrate into the tissues is a crucial factor. In the case when the degree of protein binding is high, e.g. b-lactams, changes in the free versus bound concentration not only affect the distributional characteristics of the antibiotic itself but also its level of antimicrobial activity as the free form is commonly considered to be the active moiety [18,19]. Metabolism and elimination, on the other hand, act as opposing factors against systemic exposure and their rates directly affect the dosage regimen required. Relevant factors that exist between antibiotic dosing and expression of antimicrobial effects are given in Fig. 2.
6. In vitro systems To evaluate more efficiently the impact of pharmacokinetics on microbial pharmacodynamics, various in vitro pharmacokinetic methods have been developed and their applications have been thoroughly discussed in the literature [20–23]. A simple one-compartment in
232
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
vitro model operated by continuous dilution was first introduced by Grasso [20]. To date, much more advanced models are available for complicated studies involving tissue distribution kinetics [20,24,25], antibiotic combinations with different elimination half-lives [26,27] and different modes of drug input [20,26–29], etc. Unlike in vivo conditions where pharmacokinetic parameters are interrelated, e.g. increases in Cmax and AUC are proportional in a linear system, the contribution of each pharmacokinetic parameter to the microbial dynamics can now be varied and tested independently. More importantly, the flexibility of these models allows changes in various pharmacokinetic and pharmacodynamic parameters, e.g. peak/MIC, T \ MIC, AUC \ MIC, AUIC, to be studied among antibiotics of the same class and also between different classes such that the most appropriate parameter describing the antimicrobial responses can be easily identified. Although these in vitro kinetic models are attractive in terms of providing a more realistic environment when dynamic changes of drug concentrations in vivo can be reproduced, some limitations do exist with their use. Among other factors such as a faster rate of cell growth, a different colonization pattern, e.g. isolated cells versus film/vegetation, and a lower degree of virulence in vitro, the lack of all relevant in vivo factors including cellular and humoral immune responses is most critical [30]. As a result, overgrowth of the organism at later times during a study can commonly be observed when in vitro systems are used; this is especially more so for less potent antibiotics. On the other hand, complete eradication can be a problem for potent antibiotics exhibiting a long half-life. However, when properly employed, these in vitro systems do allow the
intrinsic antimicrobial activity [24] and emergence of bacterial adaptation [31] to be examined.
7. Development of recent pharmacokinetic/pharmacodynamic markers One of the ultimate goals of utilizing pharmacokinetic/pharmacodynamic principles in the evaluation of antimicrobial responses is to improve the selection of dosage regimen. However, most data reported in the literature are based on those obtained from in vitro and/or animal studies [32–37]. In addition to the limitations of such in vitro systems discussed previously, pharmacokinetic differences between animals and humans [38] render data extrapolation from these preclinical studies to the actual clinical situations rather difficult. A more intensive search for ways to allometrically scale preclinical data becomes necessary. Nevertheless, success in the application of some of these in vitro/preclinical data to humans has been documented. The idea of maintaining plasma antibiotic concentrations above the MIC for b-lactams is based on the observation that a maximum rate of bacterial killing can be achieved beyond a certain concentration close to the MIC, i.e. concentration independent killing [2,3,38,39]. In clear contrast, the bactericidal effect exhibited by the aminoglycosides and quinolones shows a low degree of saturability with concentrations, i.e. concentration dependent killing [38,39]. In this case, higher concentrations should result in a parallel increase in antimicrobial effects. Based on these observations, the time period when plasma concentrations exceed the MIC or the area under the curve above the MIC (AUC \ MIC) become relevant gauges for therapeutic success for antibiotics exhibiting concentration independent killing, (Fig. 3) [39–41]. As for antibiotics showing concentration dependent killing, the peak to MIC or AUC to MIC ratios are more relevant indexes [11,39– 42]. Therefore, a good understanding of the pharmacological behaviour of the antibiotic in relation to its pharmacokinetics is an important factor in the quest for a better dosing regimen.
8. Concentration/dose/exposure –response relationship
Fig. 3. Some common pharmacokinetic/pharmacodynamic surrogate markers used in relating the expression of antimicrobial activity; peak plasma concentration for the estimation of peak to MIC ratio, AUC \ MIC, i.e. area above the MIC, and time above MIC.
When evaluating the concentration dependence of bacterial killing for different classes of antibiotics, it is conventional that a nonlinear relationship between bactericidal rate constants and concentrations is observed [2,3,5,7,8,43]. Via a mathematical modeling approach, a sigmoidal relationship between these two variables can be defined and is represented by the equation shown below.
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
E =Emax C S/(ECS50 +C S) where E is the rate of bacterial killing observed at concentration C, Emax is the maximal rate of bacterial killing attainable by that agent, EC50 is the concentration at which 50% of the Emax is achieved and S is the slope factor describing the steepness of the concentration-response curve. Interestingly, this relationship does not only describe bactericidal rate constant but can be readily extended to other pharmacological effects including the PAE [16,17]. A similar relationship is applicable to clinical endpoints including the rate of bacterial eradication, time to resolution of clinical symptoms etc. (as a dependant variable) and by replacing concentration with a pharmacokinetic/ pharmacodynamic marker, e.g. AUC \ MIC, T \ MIC, AUIC, etc. (as an independent variable). By establishing such a relationship, clinical outcome/success can be more precisely predicted based on both available pharmacokinetic and pharmacodynamic information.
9. A universal pharmacokinetic/pharmacodynamic marker Because of the availability of a number of pharmacokinetic/pharmacodynamic markers for use in describing different classes of antibiotics, it is prudent that, if possible, a relevant marker can be derived for efficacy estimation across different antibiotic classes. Recently, it has been proposed that the 24 h area under the inhibitory curve (AUIC) can potentially be used to serve for such purpose [41,44,45]. By convention, the AUIC is defined as the AUC/MIC normalized for a 24 h period with the intent to accommodate the various dosing regimens required for different antibiotics. For three antibiotics representing the b-lactam, quinolone and aminoglycoside classes, a similar cutoff value of 125 SIT − 1 was derived via mathematical simulations such that all three antibiotics are considered to be equi-efficacious [44]. The clinical utility of this approach has been verified in a recent study with ciprofloxacin in seriously ill patients [46]. Indeed, this approach represents a step forward such that all pharmacokinetic and pharmacodynamic determinants are taken into consideration by a single marker for different antibiotics.
10. Resistance and microbial pharmacodynamics Microbial resistance is a major factor affecting the final outcome of any antimicrobial chemotherapy. Assuming that the selection of antibiotic is appropriate with respect to the available susceptibility informa-
233
tion, two major conditions may lead to the development of resistance, i.e. a poor immune function of the host and/or inappropriate dosage regimen. From the experiences obtained in AIDS or cancer patients, deficiency of the immune status is an intrinsic limitation to the success of antimicrobial chemotherapy. To this end, the primary goal of antibiotic use is to secure a high level of bactericidal activity in an attempt to alleviate clinical symptoms and to delay resistance formation. In managing patients with intact immune function, the employment of an appropriate dosage regimen is more relevant for preventing resistance development. In estimating the potential of resistance development, characterization of the mutation frequency in relation to antibiotic concentrations or the shift in susceptibility in relation to the extent/time of clinical use of the antibiotic provides some preliminary information [47,48]. However, resistance induction potentials can be obtained by submitting the organism to serial daily passages of the test antibiotic. Recent studies on Enterobacter cloacae [49,50] and Citrobacter freundii [51] showed an inverse relationship between antibiotic potency and resistance induction. In order words, highly potent antibiotics not only invoke rapid bacterial eradication but minimize the probability of resistance formation at the same time. Therefore, potent antibiotics with low intrinsic resistance potentials should be reserved as last line agents in an attempt to conserve the effectiveness of these antibiotics. From a clinical point of view, adequate knowledge of the dose requirement and dosing regimen for individual antibiotics is important in minimizing resistance formation. Because of the apparent inverse association between a positive clinical response and the development of resistance, a recent study has demonstrated that by attaining an AUIC of \ 100 for any single antibiotic or by employing antibiotic combinations, the rates of resistance formation could be lowered [52]. Therefore, proper employment of pharmacokinetic/pharmacodynamic markers not only maximizes the antimicrobial effects but also reduces the emergency of resistant subpopulations [53].
11. Conclusions Over recent years, the use of surrogate markers in antimicrobial chemotherapy has led to more cost-effective use of antibiotics. This is a direct consequence of improved understanding of the pharmacological, pharmacokinetic and pharmacodynamic characteristics of the antibiotic and other host related factors. Attempts towards better integration of all these factors not only improve clinical use of antibiotics but result in better prediction of therapeutic outcomes.
234
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
References [1] National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. 3rd ed. Approved standard. NCCLS document M7-A3. National Committee for Clinical Laboratory Standards, Villanova, PA, 1993. [2] Li RC, Nix DE, Schentag JJ. Pharmacodynamic modeling of bacterial kinetics: b-lactam antibiotics against Escherichia coli. J Pharm Sci 1994;83:970–5. [3] Li RC. Simultaneous pharmacodynamic analysis of the lag and bactericidal phases exhibited by b-lactams against Escherichia coli. Antimicrob Agents Chemother 1996;40:2306–10. [4] Hamano S, Tsuji A, Asano T, et al. Kinetic analysis and characterization of the bacterial regrowth after treatment of Escherichia coli with b-lactam antibiotics. J Pharm Sci 1984;7310:1422– 7. [5] Tsuji A, Hamano S, Asano T, et al. Microbial kinetics of b-lactam antibiotics against Escherichia coli. J Pharm Sci 1984;73(10):1418– 22. [6] Ma HHM, Chiu FCK, Li RC. Mechanistic investigation of the reduction in antimicrobial activity of ciprofloxacin by metal cations. Pharm Res 1997;14:366–70. [7] Li RC, Schentag JJ, Nix DE. The FME method: a new way to characterize the effect of antibiotic combinations and other nonlinear pharmacologic responses. Antimicrob Agents Chemother 1993;37:523–31. [8] Li RC, Nix DE, Schentag JJ. Performance of the fractional maximal effect method: comparative interaction studies of ciprofloxacin and protein synthesis inhibitors. J Chemother 1996;8:25 – 32. [9] Craig WA, Gudmundsson S. The postantibiotic effect. In: Lorian V, editor. Antibiotics in Laboratory Medicine. Baltimore: William & Wilkins Co, 1991:403–31. [10] Zhanel GC, Hoban DJ, Harding GKM. The postantibiotic effect: a review of in vitro and in vivo data. DICP Ann Pharmacother 1991;25:153 – 63. [11] Lacy MK, Nicolau DP, Nightingale CH, Quintiliani R. The pharmacodynamics of aminoglycosides. Clin Infect Dis 1998;27:23 – 7. [12] DeVries PJ, Verkooyen RP, Leguit P, et al. Prospective randomized study of once-daily versus thrice daily netilmicin regimens in patients with intraabdominal infection. Eur J Clin Microbiol Infect Dis 1990;9:161–8. [13] Gould IM. Pharmacodynamics and the relationship between in vitro and in vivo activity of antimicrobial agents. J Chemother 1997;9:74 – 83. [14] McDonald PJ, Wetherall BL, Pruul H. Postantibiotic leukocyte enhancement: increased susceptibility of bacteria pretreated with antibiotics to activity of leukocyte. Rev Infect Dis 1981;3:38 – 44. [15] Li RC, Zhu ZY, Lee SW, Raymond K., Ling JML, Augustine FB Cheng. Antibiotic exposure and its relationship to postantibiotic effect and bactericidal activity: constant versus exponentially decreasing tobramycin levels against Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 1997; 41:1808–11. [16] Zhu ZY, Li RC. Impact of pharmacokinetics on the postantibiotic effect exhibited by Pseudomonas aeruginosa following tobramycin exposure: application of an in vitro model. J Antimicrob Chemother 1998;42:61–5. [17] Li RC, Lee SW, Kong CH. Correlation between bactericidal activity and postantibiotic effect for five antibiotics with different mechanisms of action. J Antimicrob Chemother 1997;40:39 – 45. [18] Peterson LR, Moody JA. Fasching. Influence of protein binding on therapeutic efficacy of cefoperazone. Antimicrob Agents Chemother 1989;33:566–8.
[19] Li RC, Cheng NC, Yung L. Impact of protein binding on antimicrobial activity: application of the gradient plate technique on ceftriaxone, In: Annual Meeting of the American Association of Pharmaceutical Scientists: 1997 Nov 2 – 6. Boston Pharm Res 1997;14:S71. [20] Grasso S, Meinardi G, Carneri I, Tamassia V. New in vitro model to study the effect of antibiotic concentration and rate of elimination on antibacterial activity. Antimicrob Agents Chemother 1978;13:570 – 6. [21] Nishida M, Murakawa T, Kamimura T, Okada T. Bactericidal activity of cephalosporins in an in vitro model simulating serum levels. Antimicrob Agents Chemother 1978;6:6 – 12. [22] Bergan T, Carlsen IB, Fuglesang JE. An in vitro model for monitoring bacterial responses to antibiotic agents under simulated in vivo conditions. Infection 1980;8:S96 – S102. [23] Grasso S. Historical review of in-vitro models. J Antimicrob Chemother (Suppl) 1985;15:99 – 102. [24] Reeves DS. Advantages and disadvantages of an in vitro model with two compartments connected by a dialyser: results of experiments with ciprofloxacin. J Antimicrob Chemother (Suppl) 1985;15:159 – 67. [25] Blaser J, Stone BB, Zinner SH. Two compartment kinetic model with multiple artificial capillary units. J Antimicrob Chemother Suppl 1985;15:131 – 7. [26] Blaser J. In-vitro model for simultaneous simulation of the serum kinetics of two drugs with different half-lives. J Antimicrob Chemother (Suppl) 1985;15:125 – 30. [27] Shah PM. Simultaneous simulation of two different concentration time curves in vitro. J Antimicrob Chemother (Suppl) 1985;15:261 – 4. [28] Bergan T, Carlsen IB. Effect of antibiotics eliminated by first order kinetics. J Antimicrob Chemother (Suppl) 1985;15:147–52. [29] Zabinski RA, Vance-Bryan K, Krinke AJ, et al. Evaluation of activity of temafloxacin against Bacteroides fragilis by an in vitro pharmacodynamic system. Antimicrob Agents Chemother 1993;37:2454 – 8. [30] Dalhoff A. Differences between bacteria grown in vitro and in vivo. J Antimicrob Chemother (suppl) 1985;15:175 – 95. [31] Seeberg AH, Wiedeman B. Application of in vitro models: development of resistance. J Antimicrob Chemother (Suppl) 1985;15:241 – 9. [32] Nolting A, Dalla Costa T, Rand KH, et al. Pharmacokineticpharmacodynamic modeling of the antibiotic effect of pipericillin in vitro. Pharm Res 1996;13(1):91 – 6. [33] Madaras-Kelly KJ, Ostergaard BE, Havde LB, et al. Twentyfour-hour area under the concentration-time curve/MIC ratio as a generic predictor of fluroquinolone antimicrobial effect by using three strains of Pseudomonas aeruginosa and an in vitro pharmacodynamic model. Antimicrob Agents Chemother 1996;40(3):627– 32. [34] Chavanet P, Dalle F, Delisle P, et al. Experimental efficacy of combined ceftriaxone and amoxycillin on penicillin-resistant and broad-spectrum cephalosporin-resistant Streptococcus pneumoniae infection. J Antimicrob Chemother 1998;41:237 – 46. [35] Zhi JG, Nightingale CH, Quintiliani R. Microbial pharmacodynamics of piperacillin in neutropenic mice of systematic infection due to Pseudomonas aeruginosa. J Pharmacokinet Biopharm 1988;16:355 – 75. [36] Johnson DE, Thompson B, Calia FM. Comparative activities of piperacillin, ceftazidime, and amikacin, alone and in all possible combinations, against experimental Pseudomonas aeruginosa infections in neutropenic rats. Antimicrob Agents Chemother 1985;28:735 – 9. [37] Craig WA, Redington J, Ebert SC. Pharmacodynamics of amikacin in vitro and in mouse thigh and lung infections. J Antimicrob Chemother (Suppl C) 1991;27:29 – 40.
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235 [38] Vogelman B, Gudmundsson S, Leggett J, et al. Correlation of antimirobial pharmacokinetic paramteres with therapeutic efficacy in an animal model. J Infect Dis 1988;158(4):831–47. [39] Craig WA. Pharmacokinetic/pharmacodynamic parameters: rationale for antibacterial dosing of mice and men. Clin Infect Dis 1998;26:1 – 12. [40] Nicolau DP, Quintiliani R, Nightingale CH. Antibiotic kinetics and dynamic for the clinican. Med Clin N Am 1995;79(3):477– 95. [41] Hyatt JM, McKinnon PS, Zimmer GS, et al. The importance of pharmacokinetic/pharmacodynamic surrogate markers to outcome: focus on antibacterial agents. Clin Pharmacokinet 1995;28(2):143– 60. [42] Moore RD, Lietman PS, Smith CR. Clinical response to aminoglycoside therapy: importance of the ratio of peak concentration to minimal inhibitory concentration. J Infect Dis 1987;155(1):93– 9. [43] Mattie H. A mathematical description of short-term effects of b-lactam antibiotics on bacterial growth in vitro. Cur Micobiol 1978;1:106 – 9. [44] Schentag JJ, Nix DE, Adelman MH. Mathematical examination of dual individualization principles (I): relationships between AUC above MIC and area under the inhibitory curve for cefmenoxime, ciprofloxacin, and tobramycin. DICP 1991;25(10):1050– 7. [45] Schentag JJ, Nix DE, Forrest A, Adelman MH. AUIC – the universal parameter within the constraint of a reasonable dosing interval. Ann Pharmacother 1996;30(9):1029–31.
.
235
[46] Forrest A, Nix DE, Ballow CH, Pharmacodynamics of intravenous ciprofloxacin in seriously ill patients. Antimicrob. Agents Chemother, 1993; 1073 – 81. [47] McGowan JE Jr., Antimicrobial resistance in hospital organisms and its relation to antibiotic use. Rev Infect Dis 1983;5:1033-48. [48] Gaynes RP, Culver DH, Horan TC, et al. Trends in methicillinresistant Staphylococcus aureus in United States hospitals. Infect Dis Clin Prac 1992;2:452 – 5. [49] Chan WC, Li RC, Ling JM, et al. Markedly different rates and resistance profiles exhibited by seven commonly used and newer beta-lactams on the selection of resistant variants of Enterobacter cloacae. J Antimicrob Chemother 1999;43:55 – 60. [50] Fung-Tomc JC, Gradelski E, Huczko E, et al. Differences in the resistant variants of Enterobacter cloacae selected by extendedspectrum cephalosporins. Antimicrob Agents Chemother 1996;40:1289 – 93. [51] Stapleton P, Shannon K, Phillips I. The ability of b-lactam antibiotics to select mutants with depressed b-lactamase synthesis from Citrobacter freundii. J Antimicrob Chemother 1995;36:483 – 96. [52] Thomas JK, Forrest A, Bhavnani SM, et al. Pharmacodynamic evaluation of factors associated with the development of bacterial resistance in acutely ill patients during therapy. Antimicrob Agents Chemother 1998;42(3):521– 7. [53] Schentag JJ, Birmingham MC, Paladino JA, et al. In nosocomial pneumonia, optimizing antibiotics other than aminoglycosides is a more important determinant of successful clinical outcome, and a better means of avoiding resistance. Sem Respir Infect 1997;12(4):278– 93.
www.ischemo.org
Review
New pharmacodynamic parameters for antimicrobial agents Ronald C. Li * Pharmacokinetic/Pharmacodynamic Sciences, Genetics Institute, One Burtt Road, Ando6er, MA 01810, USA Received 11 May 1999; accepted 14 May 1999
Abstract The application of pharmacodynamic theories to antimicrobial chemotherapy has greatly improved the prediction of the time course of activity expressed by antibiotics. Being a major component of the antibiotic – bacterium interaction system, pharmacodynamics, when properly integrated with the pharmacokinetics established for the antibiotic, allow better evaluation of the dosage regimen in conjunction with its clinical response. Before this approach becomes effective, detailed background information on the complex antibiotic–bacterium interactions have to be secured. To achieve this, proper characterization of a time – kill curve is a prerequisite. The use of susceptibility endpoints such as the MIC with respect to the antibiotic concentrations achievable in vivo represent the conventional approach to clinical dosing of antimicrobial agents, i.e. by maintaining concentrations above the MIC. Recently, a number of surrogate markers have been proposed by combining suitable pharmacokinetic parameters and susceptibility data, e.g. peak/MIC ratio, AUC \ MIC, time above MIC, AUIC etc. to enhance the prediction of clinical outcomes. Attempts have been made to apply these pharmacokinetic/pharmacodynamic markers to antibiotics of the same class as well as to antibiotics from different classes. This review aims to discuss the various microbial dynamic responses in relation to antibiotic exposure and the development of different pharmacokinetic/pharmacodynamic markers for use in current antimicrobial chemotherapy. © 2000 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. Keywords: Antibiotic – bacterium interaction system; Time–kill curve; Pharmacokinetic factors
1. Introduction Over the past 2 to 3 decades, there has been a steady yet persistent increase in the use of pharmacodynamics to describe the time course of antimicrobial activity exhibited by various antibiotics. Part of the reason is the need to improve accuracy for delineating the complex interactions that exist between the antibiotic and the pathogen. As antibiotics are not instant poisons, the sequence of events demonstrating the action(s) of the antibiotic and the reaction(s) of the microorganisms usually extends over a period of time. To appreciate fully the complexity of the interactions, understanding the fundamental elements in relation to time is therefore vital not only from the perspective of evaluating the activity profiles of currently available antibiotics * Corresponding author. Tel.: +1-978-2471884; fax: + 1-9782471389. E-mail address: [email protected] (R.C. Li)
but the design of newer ones. In addition, it has always been the intent of researchers as well as clinicians to derive and extrapolate pharmacodynamic parameters in order to predict therapeutic responses more accurately.
2. Routine susceptibility testing The minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) are conventional measures to aid proper antibiotic selection for clinical use. The MIC measurement involves subjective determination of turbidity after the isolated pathogen is exposed to graded two-fold dilutions of the antibiotic for 18–24 h at 35–37°C in broth [1] whereas, MBC is the lowest concentration that produces a \99.99% reduction in viable bacterial counts in comparison to that prior to antibiotic exposure. It is apparent that these two measurements provide a single time point
0924-8579/00/$20 © 2000 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. PII: S 0 9 2 4 - 8 5 7 9 ( 9 9 ) 0 0 1 1 4 - 4
230
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
observation with no information on the temporal changes in viable bacterial cell density over the incubation period. As expected, functional use of these measurements in defining microbial pharmacodynamics is minimal.
3. Time–kill curves: pharmacodynamics and basic pharmacology The diversity of various bacterial responses to antibiotics cannot be easily understood without the establishment of a time–kill curve. A typical time – kill curve generated by subjecting the test organism to exposure at constant antibiotic concentrations, normally consists of a lag phase, a log-linear killing phase a second lag phase and a regrowth phase [2 – 5]. A static phase can be observed when resistant bacterial subpopulations overgrow to a saturated level when nutrients in the medium are exhausted. Understanding the sequence of events that takes place between the antibiotic and bacteria is vital in order to interpret the dynamics of these individual phases. Entry of the antibiotic into bacterial cells is an important first step for subsequent expression of antimicrobial activity. The cell wall/membrane usually offers an initial defensive mechanism against antibiotic permeation. For a
Fig. 1. Impact of the decreasing rate of resistance formation on the time–kill curve. When the rate of bactericidal effect remained identical at 5 h − 1 for all simulations, the reduction in the rate of resistance formation (values shown in graph) caused a delay in the development and overgrowth of resistant subpopulations. Data adapted from [2].
rapidly bactericidal antibiotic, permeation is probably a rate-limiting step for the expression of cell death. For less bactericidal antibiotics, the time lag observed between initiation of antibiotic exposure and commencement of bacterial killing may represent, in addition to permeation, the time required for all the intermediate steps that take place between receptor binding and cell death [2,3,5]. Regardless of the exact limiting factor, a high antibiotic concentration is an overriding force for shortening this lag by producing a concentration gradient favourable for cellular permeation as well as equilibration towards higher receptor occupancy. From a pharmacological perspective, the log-linear killing phase has been considered to be the most relevant indicator for bactericidal activity, i.e. as a quantitative measure for the intensity of bactericidal effect. Due to the fact that bacterial killing is the ultimate consequence following binding of the antibiotic molecules to their intended receptors and in line with most pharmacological effects, the bactericidal rate constants are saturable at high concentrations [2,3]. Generation of the concentration-response relationships between different antibiotics sharing the same mechanism of action allows direct comparison of their intrinsic potencies [2,3,5–8]. Similarly, bactericidal rate constants also permit the decrease of viable cell counts in the killing phase to be more precisely predicted during antibiotic exposure. In comparison, the regrowth phase has not been studied as extensively. Although bacterial regrowth is not a desirable effect during antibiotic exposure, it does reflect the ability of the organism to adapt to the antibiotic assault. For a number of b-lactams, the rate of bacterial resistance formation has been shown to reduce exponentially at increasing antibiotic concentrations [2]. In other words, the probability for resistant subpopulations to develop and overgrow is drastically reduced at higher antibiotic concentrations. Antibiotics showing a low potential of resistance formation should allow the bactericidal phase to be extended over a longer time for more extensive cell kill (Fig. 1). Therefore, the establishment of a full pharmacodynamic profile over time can be very informative for evaluation of the different aspects of antibiotic activity and the dynamics associated with the antibiotic-bacterium interaction. In principle, proper interpretation of various parameters describing the time–kill curve should also effectively enhance the design of new antibiotics in terms of cell permeation and receptor binding as well as resistance formation.
4. Beyond time–kill curve: postantibiotic antibiotic effect Characterization of a time–kill curve has not satisfied the quest for more endpoints describing other
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
231
Fig. 2. Antibiotic-bacterium-host interactions that exist between administration of a dose of antibiotic and the expression of antimicrobial activity.
important components of the microbial dynamic system. The persistent suppression of the bacterial growth which follows antibiotic exposure is now known as the postantibiotic effect (PAE) and is one of the more relevant pharmacodynamic parameters with respect to the optimization of dosage regimen. The PAE is highly dependent on the antibiotic – bacterial combination under investigation [9,10]. Theoretically, PAE offers a good argument for extending the dosing interval for any antibiotic exhibiting such effect. This interpretation is based on the fact that growth suppression remains when antibiotic concentration fell below its respective MIC value. Indeed, clinical utilization of PAE has been shown to be successful in aminoglycoside therapy by replacing the traditional divided daily dosing with the high dose once daily dosing regimen [11,12]. Other phenomena relating to PAE which include growth suppression following antibiotic exposure at subinhibitory concentrations (PAESME) and the enhancement of phagocytic activity during the PAE phase, also point to the functionality of PAE in designing dosage regimen for antibiotics [13,14]. The inclusion of PAE in the design of a dosing regimen cannot be completed without the consideration of pharmacokinetics as an integral factor. A recent study has demonstrated that the PAE induced is relatively longer for organisms exposed to constant antibiotic concentrations rather than exponentially decreasing concentrations even at a similar degree of antibiotic exposure in terms of AUC \ MIC [15]. This observation suggests the expression of PAE is dependent not only on the magnitude but also the mode of antibiotic exposure. Data collected from a subsequent study investigating the length of dosing interval as an independent variable suggest the state of cellular damage/recovery is a relevant determinant for the PAE observed [16,17]. These previous observations on PAE clearly demonstrated the contribution of pharmacokinetics to microbial pharmacodynamics. Unequivocally, the impact of pharmacokinetics on pharmacodynamics should be clearly defined in order to permit more accurate extrapolation of in vitro data. However, for most pharmacodynamic studies, the constraint of em-
ploying a static environment of constant antibiotic concentrations remains.
5. Pharmacokinetics: a determinant for pharmacodynamics in antimicrobial chemotherapy Pharmacokinetics are tools to depict the fundamental processes of absorption, distribution, metabolism and elimination (ADME) of any xenobiotics administered to a biological system. From the perspective of antimicrobial therapy, these processes in turn dictate the concentration versus time profile of an antibiotic and its activity in vivo. For any antibiotic to express its activity, absorption is a prerequisite to achieve adequate systemic presence in relation to the level of susceptibility. After confirming an adequate degree of absorption, the most basic approach in early day antimicrobial chemotherapy has been simply to maintain systemic levels above the MIC by varying the dose and/or dosing regimen. As for the treatment of deep tissue infections, the ability for the antibiotic to distribute or penetrate into the tissues is a crucial factor. In the case when the degree of protein binding is high, e.g. b-lactams, changes in the free versus bound concentration not only affect the distributional characteristics of the antibiotic itself but also its level of antimicrobial activity as the free form is commonly considered to be the active moiety [18,19]. Metabolism and elimination, on the other hand, act as opposing factors against systemic exposure and their rates directly affect the dosage regimen required. Relevant factors that exist between antibiotic dosing and expression of antimicrobial effects are given in Fig. 2.
6. In vitro systems To evaluate more efficiently the impact of pharmacokinetics on microbial pharmacodynamics, various in vitro pharmacokinetic methods have been developed and their applications have been thoroughly discussed in the literature [20–23]. A simple one-compartment in
232
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
vitro model operated by continuous dilution was first introduced by Grasso [20]. To date, much more advanced models are available for complicated studies involving tissue distribution kinetics [20,24,25], antibiotic combinations with different elimination half-lives [26,27] and different modes of drug input [20,26–29], etc. Unlike in vivo conditions where pharmacokinetic parameters are interrelated, e.g. increases in Cmax and AUC are proportional in a linear system, the contribution of each pharmacokinetic parameter to the microbial dynamics can now be varied and tested independently. More importantly, the flexibility of these models allows changes in various pharmacokinetic and pharmacodynamic parameters, e.g. peak/MIC, T \ MIC, AUC \ MIC, AUIC, to be studied among antibiotics of the same class and also between different classes such that the most appropriate parameter describing the antimicrobial responses can be easily identified. Although these in vitro kinetic models are attractive in terms of providing a more realistic environment when dynamic changes of drug concentrations in vivo can be reproduced, some limitations do exist with their use. Among other factors such as a faster rate of cell growth, a different colonization pattern, e.g. isolated cells versus film/vegetation, and a lower degree of virulence in vitro, the lack of all relevant in vivo factors including cellular and humoral immune responses is most critical [30]. As a result, overgrowth of the organism at later times during a study can commonly be observed when in vitro systems are used; this is especially more so for less potent antibiotics. On the other hand, complete eradication can be a problem for potent antibiotics exhibiting a long half-life. However, when properly employed, these in vitro systems do allow the
intrinsic antimicrobial activity [24] and emergence of bacterial adaptation [31] to be examined.
7. Development of recent pharmacokinetic/pharmacodynamic markers One of the ultimate goals of utilizing pharmacokinetic/pharmacodynamic principles in the evaluation of antimicrobial responses is to improve the selection of dosage regimen. However, most data reported in the literature are based on those obtained from in vitro and/or animal studies [32–37]. In addition to the limitations of such in vitro systems discussed previously, pharmacokinetic differences between animals and humans [38] render data extrapolation from these preclinical studies to the actual clinical situations rather difficult. A more intensive search for ways to allometrically scale preclinical data becomes necessary. Nevertheless, success in the application of some of these in vitro/preclinical data to humans has been documented. The idea of maintaining plasma antibiotic concentrations above the MIC for b-lactams is based on the observation that a maximum rate of bacterial killing can be achieved beyond a certain concentration close to the MIC, i.e. concentration independent killing [2,3,38,39]. In clear contrast, the bactericidal effect exhibited by the aminoglycosides and quinolones shows a low degree of saturability with concentrations, i.e. concentration dependent killing [38,39]. In this case, higher concentrations should result in a parallel increase in antimicrobial effects. Based on these observations, the time period when plasma concentrations exceed the MIC or the area under the curve above the MIC (AUC \ MIC) become relevant gauges for therapeutic success for antibiotics exhibiting concentration independent killing, (Fig. 3) [39–41]. As for antibiotics showing concentration dependent killing, the peak to MIC or AUC to MIC ratios are more relevant indexes [11,39– 42]. Therefore, a good understanding of the pharmacological behaviour of the antibiotic in relation to its pharmacokinetics is an important factor in the quest for a better dosing regimen.
8. Concentration/dose/exposure –response relationship
Fig. 3. Some common pharmacokinetic/pharmacodynamic surrogate markers used in relating the expression of antimicrobial activity; peak plasma concentration for the estimation of peak to MIC ratio, AUC \ MIC, i.e. area above the MIC, and time above MIC.
When evaluating the concentration dependence of bacterial killing for different classes of antibiotics, it is conventional that a nonlinear relationship between bactericidal rate constants and concentrations is observed [2,3,5,7,8,43]. Via a mathematical modeling approach, a sigmoidal relationship between these two variables can be defined and is represented by the equation shown below.
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
E =Emax C S/(ECS50 +C S) where E is the rate of bacterial killing observed at concentration C, Emax is the maximal rate of bacterial killing attainable by that agent, EC50 is the concentration at which 50% of the Emax is achieved and S is the slope factor describing the steepness of the concentration-response curve. Interestingly, this relationship does not only describe bactericidal rate constant but can be readily extended to other pharmacological effects including the PAE [16,17]. A similar relationship is applicable to clinical endpoints including the rate of bacterial eradication, time to resolution of clinical symptoms etc. (as a dependant variable) and by replacing concentration with a pharmacokinetic/ pharmacodynamic marker, e.g. AUC \ MIC, T \ MIC, AUIC, etc. (as an independent variable). By establishing such a relationship, clinical outcome/success can be more precisely predicted based on both available pharmacokinetic and pharmacodynamic information.
9. A universal pharmacokinetic/pharmacodynamic marker Because of the availability of a number of pharmacokinetic/pharmacodynamic markers for use in describing different classes of antibiotics, it is prudent that, if possible, a relevant marker can be derived for efficacy estimation across different antibiotic classes. Recently, it has been proposed that the 24 h area under the inhibitory curve (AUIC) can potentially be used to serve for such purpose [41,44,45]. By convention, the AUIC is defined as the AUC/MIC normalized for a 24 h period with the intent to accommodate the various dosing regimens required for different antibiotics. For three antibiotics representing the b-lactam, quinolone and aminoglycoside classes, a similar cutoff value of 125 SIT − 1 was derived via mathematical simulations such that all three antibiotics are considered to be equi-efficacious [44]. The clinical utility of this approach has been verified in a recent study with ciprofloxacin in seriously ill patients [46]. Indeed, this approach represents a step forward such that all pharmacokinetic and pharmacodynamic determinants are taken into consideration by a single marker for different antibiotics.
10. Resistance and microbial pharmacodynamics Microbial resistance is a major factor affecting the final outcome of any antimicrobial chemotherapy. Assuming that the selection of antibiotic is appropriate with respect to the available susceptibility informa-
233
tion, two major conditions may lead to the development of resistance, i.e. a poor immune function of the host and/or inappropriate dosage regimen. From the experiences obtained in AIDS or cancer patients, deficiency of the immune status is an intrinsic limitation to the success of antimicrobial chemotherapy. To this end, the primary goal of antibiotic use is to secure a high level of bactericidal activity in an attempt to alleviate clinical symptoms and to delay resistance formation. In managing patients with intact immune function, the employment of an appropriate dosage regimen is more relevant for preventing resistance development. In estimating the potential of resistance development, characterization of the mutation frequency in relation to antibiotic concentrations or the shift in susceptibility in relation to the extent/time of clinical use of the antibiotic provides some preliminary information [47,48]. However, resistance induction potentials can be obtained by submitting the organism to serial daily passages of the test antibiotic. Recent studies on Enterobacter cloacae [49,50] and Citrobacter freundii [51] showed an inverse relationship between antibiotic potency and resistance induction. In order words, highly potent antibiotics not only invoke rapid bacterial eradication but minimize the probability of resistance formation at the same time. Therefore, potent antibiotics with low intrinsic resistance potentials should be reserved as last line agents in an attempt to conserve the effectiveness of these antibiotics. From a clinical point of view, adequate knowledge of the dose requirement and dosing regimen for individual antibiotics is important in minimizing resistance formation. Because of the apparent inverse association between a positive clinical response and the development of resistance, a recent study has demonstrated that by attaining an AUIC of \ 100 for any single antibiotic or by employing antibiotic combinations, the rates of resistance formation could be lowered [52]. Therefore, proper employment of pharmacokinetic/pharmacodynamic markers not only maximizes the antimicrobial effects but also reduces the emergency of resistant subpopulations [53].
11. Conclusions Over recent years, the use of surrogate markers in antimicrobial chemotherapy has led to more cost-effective use of antibiotics. This is a direct consequence of improved understanding of the pharmacological, pharmacokinetic and pharmacodynamic characteristics of the antibiotic and other host related factors. Attempts towards better integration of all these factors not only improve clinical use of antibiotics but result in better prediction of therapeutic outcomes.
234
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235
References [1] National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. 3rd ed. Approved standard. NCCLS document M7-A3. National Committee for Clinical Laboratory Standards, Villanova, PA, 1993. [2] Li RC, Nix DE, Schentag JJ. Pharmacodynamic modeling of bacterial kinetics: b-lactam antibiotics against Escherichia coli. J Pharm Sci 1994;83:970–5. [3] Li RC. Simultaneous pharmacodynamic analysis of the lag and bactericidal phases exhibited by b-lactams against Escherichia coli. Antimicrob Agents Chemother 1996;40:2306–10. [4] Hamano S, Tsuji A, Asano T, et al. Kinetic analysis and characterization of the bacterial regrowth after treatment of Escherichia coli with b-lactam antibiotics. J Pharm Sci 1984;7310:1422– 7. [5] Tsuji A, Hamano S, Asano T, et al. Microbial kinetics of b-lactam antibiotics against Escherichia coli. J Pharm Sci 1984;73(10):1418– 22. [6] Ma HHM, Chiu FCK, Li RC. Mechanistic investigation of the reduction in antimicrobial activity of ciprofloxacin by metal cations. Pharm Res 1997;14:366–70. [7] Li RC, Schentag JJ, Nix DE. The FME method: a new way to characterize the effect of antibiotic combinations and other nonlinear pharmacologic responses. Antimicrob Agents Chemother 1993;37:523–31. [8] Li RC, Nix DE, Schentag JJ. Performance of the fractional maximal effect method: comparative interaction studies of ciprofloxacin and protein synthesis inhibitors. J Chemother 1996;8:25 – 32. [9] Craig WA, Gudmundsson S. The postantibiotic effect. In: Lorian V, editor. Antibiotics in Laboratory Medicine. Baltimore: William & Wilkins Co, 1991:403–31. [10] Zhanel GC, Hoban DJ, Harding GKM. The postantibiotic effect: a review of in vitro and in vivo data. DICP Ann Pharmacother 1991;25:153 – 63. [11] Lacy MK, Nicolau DP, Nightingale CH, Quintiliani R. The pharmacodynamics of aminoglycosides. Clin Infect Dis 1998;27:23 – 7. [12] DeVries PJ, Verkooyen RP, Leguit P, et al. Prospective randomized study of once-daily versus thrice daily netilmicin regimens in patients with intraabdominal infection. Eur J Clin Microbiol Infect Dis 1990;9:161–8. [13] Gould IM. Pharmacodynamics and the relationship between in vitro and in vivo activity of antimicrobial agents. J Chemother 1997;9:74 – 83. [14] McDonald PJ, Wetherall BL, Pruul H. Postantibiotic leukocyte enhancement: increased susceptibility of bacteria pretreated with antibiotics to activity of leukocyte. Rev Infect Dis 1981;3:38 – 44. [15] Li RC, Zhu ZY, Lee SW, Raymond K., Ling JML, Augustine FB Cheng. Antibiotic exposure and its relationship to postantibiotic effect and bactericidal activity: constant versus exponentially decreasing tobramycin levels against Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 1997; 41:1808–11. [16] Zhu ZY, Li RC. Impact of pharmacokinetics on the postantibiotic effect exhibited by Pseudomonas aeruginosa following tobramycin exposure: application of an in vitro model. J Antimicrob Chemother 1998;42:61–5. [17] Li RC, Lee SW, Kong CH. Correlation between bactericidal activity and postantibiotic effect for five antibiotics with different mechanisms of action. J Antimicrob Chemother 1997;40:39 – 45. [18] Peterson LR, Moody JA. Fasching. Influence of protein binding on therapeutic efficacy of cefoperazone. Antimicrob Agents Chemother 1989;33:566–8.
[19] Li RC, Cheng NC, Yung L. Impact of protein binding on antimicrobial activity: application of the gradient plate technique on ceftriaxone, In: Annual Meeting of the American Association of Pharmaceutical Scientists: 1997 Nov 2 – 6. Boston Pharm Res 1997;14:S71. [20] Grasso S, Meinardi G, Carneri I, Tamassia V. New in vitro model to study the effect of antibiotic concentration and rate of elimination on antibacterial activity. Antimicrob Agents Chemother 1978;13:570 – 6. [21] Nishida M, Murakawa T, Kamimura T, Okada T. Bactericidal activity of cephalosporins in an in vitro model simulating serum levels. Antimicrob Agents Chemother 1978;6:6 – 12. [22] Bergan T, Carlsen IB, Fuglesang JE. An in vitro model for monitoring bacterial responses to antibiotic agents under simulated in vivo conditions. Infection 1980;8:S96 – S102. [23] Grasso S. Historical review of in-vitro models. J Antimicrob Chemother (Suppl) 1985;15:99 – 102. [24] Reeves DS. Advantages and disadvantages of an in vitro model with two compartments connected by a dialyser: results of experiments with ciprofloxacin. J Antimicrob Chemother (Suppl) 1985;15:159 – 67. [25] Blaser J, Stone BB, Zinner SH. Two compartment kinetic model with multiple artificial capillary units. J Antimicrob Chemother Suppl 1985;15:131 – 7. [26] Blaser J. In-vitro model for simultaneous simulation of the serum kinetics of two drugs with different half-lives. J Antimicrob Chemother (Suppl) 1985;15:125 – 30. [27] Shah PM. Simultaneous simulation of two different concentration time curves in vitro. J Antimicrob Chemother (Suppl) 1985;15:261 – 4. [28] Bergan T, Carlsen IB. Effect of antibiotics eliminated by first order kinetics. J Antimicrob Chemother (Suppl) 1985;15:147–52. [29] Zabinski RA, Vance-Bryan K, Krinke AJ, et al. Evaluation of activity of temafloxacin against Bacteroides fragilis by an in vitro pharmacodynamic system. Antimicrob Agents Chemother 1993;37:2454 – 8. [30] Dalhoff A. Differences between bacteria grown in vitro and in vivo. J Antimicrob Chemother (suppl) 1985;15:175 – 95. [31] Seeberg AH, Wiedeman B. Application of in vitro models: development of resistance. J Antimicrob Chemother (Suppl) 1985;15:241 – 9. [32] Nolting A, Dalla Costa T, Rand KH, et al. Pharmacokineticpharmacodynamic modeling of the antibiotic effect of pipericillin in vitro. Pharm Res 1996;13(1):91 – 6. [33] Madaras-Kelly KJ, Ostergaard BE, Havde LB, et al. Twentyfour-hour area under the concentration-time curve/MIC ratio as a generic predictor of fluroquinolone antimicrobial effect by using three strains of Pseudomonas aeruginosa and an in vitro pharmacodynamic model. Antimicrob Agents Chemother 1996;40(3):627– 32. [34] Chavanet P, Dalle F, Delisle P, et al. Experimental efficacy of combined ceftriaxone and amoxycillin on penicillin-resistant and broad-spectrum cephalosporin-resistant Streptococcus pneumoniae infection. J Antimicrob Chemother 1998;41:237 – 46. [35] Zhi JG, Nightingale CH, Quintiliani R. Microbial pharmacodynamics of piperacillin in neutropenic mice of systematic infection due to Pseudomonas aeruginosa. J Pharmacokinet Biopharm 1988;16:355 – 75. [36] Johnson DE, Thompson B, Calia FM. Comparative activities of piperacillin, ceftazidime, and amikacin, alone and in all possible combinations, against experimental Pseudomonas aeruginosa infections in neutropenic rats. Antimicrob Agents Chemother 1985;28:735 – 9. [37] Craig WA, Redington J, Ebert SC. Pharmacodynamics of amikacin in vitro and in mouse thigh and lung infections. J Antimicrob Chemother (Suppl C) 1991;27:29 – 40.
R.C. Li / International Journal of Antimicrobial Agents 13 (2000) 229–235 [38] Vogelman B, Gudmundsson S, Leggett J, et al. Correlation of antimirobial pharmacokinetic paramteres with therapeutic efficacy in an animal model. J Infect Dis 1988;158(4):831–47. [39] Craig WA. Pharmacokinetic/pharmacodynamic parameters: rationale for antibacterial dosing of mice and men. Clin Infect Dis 1998;26:1 – 12. [40] Nicolau DP, Quintiliani R, Nightingale CH. Antibiotic kinetics and dynamic for the clinican. Med Clin N Am 1995;79(3):477– 95. [41] Hyatt JM, McKinnon PS, Zimmer GS, et al. The importance of pharmacokinetic/pharmacodynamic surrogate markers to outcome: focus on antibacterial agents. Clin Pharmacokinet 1995;28(2):143– 60. [42] Moore RD, Lietman PS, Smith CR. Clinical response to aminoglycoside therapy: importance of the ratio of peak concentration to minimal inhibitory concentration. J Infect Dis 1987;155(1):93– 9. [43] Mattie H. A mathematical description of short-term effects of b-lactam antibiotics on bacterial growth in vitro. Cur Micobiol 1978;1:106 – 9. [44] Schentag JJ, Nix DE, Adelman MH. Mathematical examination of dual individualization principles (I): relationships between AUC above MIC and area under the inhibitory curve for cefmenoxime, ciprofloxacin, and tobramycin. DICP 1991;25(10):1050– 7. [45] Schentag JJ, Nix DE, Forrest A, Adelman MH. AUIC – the universal parameter within the constraint of a reasonable dosing interval. Ann Pharmacother 1996;30(9):1029–31.
.
235
[46] Forrest A, Nix DE, Ballow CH, Pharmacodynamics of intravenous ciprofloxacin in seriously ill patients. Antimicrob. Agents Chemother, 1993; 1073 – 81. [47] McGowan JE Jr., Antimicrobial resistance in hospital organisms and its relation to antibiotic use. Rev Infect Dis 1983;5:1033-48. [48] Gaynes RP, Culver DH, Horan TC, et al. Trends in methicillinresistant Staphylococcus aureus in United States hospitals. Infect Dis Clin Prac 1992;2:452 – 5. [49] Chan WC, Li RC, Ling JM, et al. Markedly different rates and resistance profiles exhibited by seven commonly used and newer beta-lactams on the selection of resistant variants of Enterobacter cloacae. J Antimicrob Chemother 1999;43:55 – 60. [50] Fung-Tomc JC, Gradelski E, Huczko E, et al. Differences in the resistant variants of Enterobacter cloacae selected by extendedspectrum cephalosporins. Antimicrob Agents Chemother 1996;40:1289 – 93. [51] Stapleton P, Shannon K, Phillips I. The ability of b-lactam antibiotics to select mutants with depressed b-lactamase synthesis from Citrobacter freundii. J Antimicrob Chemother 1995;36:483 – 96. [52] Thomas JK, Forrest A, Bhavnani SM, et al. Pharmacodynamic evaluation of factors associated with the development of bacterial resistance in acutely ill patients during therapy. Antimicrob Agents Chemother 1998;42(3):521– 7. [53] Schentag JJ, Birmingham MC, Paladino JA, et al. In nosocomial pneumonia, optimizing antibiotics other than aminoglycosides is a more important determinant of successful clinical outcome, and a better means of avoiding resistance. Sem Respir Infect 1997;12(4):278– 93.