|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 PHARMACOKINETICS AND PHARMACO DYNAMICS OF OXAZOLIDINONES AND BETA-LACTAMS By STEPHAN SCHMIDT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008
2 2008 Stephan Schmidt
3 To my family
4 ACKNOWLEDGMENTS First and forem ost, I would like to thank my supervisor Dr. Hartmut Derendorf for having my as a student, for his continuous support, advice and in times patience. Furthermore, I would also like to thank my committee members Dr. Kenneth H. Rand, Dr. Guenther Hochhaus and Dr. Veronika Butterweck for their constr uctive guidance of my thesis work. I would like to thank ou r current office staff Robin KeirnanSanchez, Patricia J. Khan and Marty Rhoden, as well as, our former office st aff Andrea Tucker and Jim Ketcham for their consistently excellent support with administrative issues during my stay in the Department of Pharmaceutics. I wish to give a special thanks to Dr. Olaf Burkhardt, April Barbour and Martina Sahre, Dr. Martin Brunner, Yufei Tang, Dr. Maria Grant, Dr Christoph N. Seubert, Dr Robert W. Rout and Dr. Kfir Ben-David for their close collaboration, our study subjects for part icipating and Shands GCRC staff for their excellent support of our clinical microdialysis projects. I would like to thank my co-w orkers Martina Sahre, April Barbour, Oliver Ghobrial and Dr. Sreedharan Nair Sabarinath, as well as, our German exchange students Katharina Roeck, Sandra Weiss, Matthias Fueth and Sara Dizayee for their support with my in vitro experiments. I want to acknowledge my co-authors April Ba rbour, Dr. Olaf Burkhardt, Dr. Kenneth H. Rand, Dr. Yanyun Li, Dr. Vipul Kumar, Dr. Edga r L. Schuck, Dr. Martin Brunner, Dr. Maria Grant, Dr. Robert W. Rout, Dr. Kfir Ben-Davi d, Dr. Seubert, Martina Sahre, Katharina Roeck, Rebecca Banks, Yufei Tang, Lars Schiefelbein and Benjamin Ma for their contributions to the manuscripts and publications. I also would like to thank Dr. Derendorfs group for the all inspiring discussions, as well as, social gatherings in and outside the laborat ory. They all have substantially improved the quality of my work and made my tim e in Gainesville a pleasant memory.
5 Last but not least, I w ould like to thank my family and frie nds, especially my future wife Arielle Pandolph, for all their support and encourag ement during my stay at the University of Florida. Chapter 2 of this thesis has been previous ly published in Antimicrobial Agents and Chemotherapy, chapter 3 in the Jo urnal of Clinical Pharmacology, chapter 4 in the Journal of Antimicrobial Agents and Chemotherapy and chapter 5 in Expert Opinion in Drug Discovery. Chapter 6 is currently under submission with Antimicrobial Agents and Chemotherapy. These journals have been credited in the respective chapters and permissions for reprint for educational purposes as part of this th esis have been granted.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4 LIST OF TABLES ...........................................................................................................................9 LIST OF FIGURES .......................................................................................................................10 LIST OF ABBREVIATIONS ........................................................................................................ 12 ABSTRACT ...................................................................................................................... .............17 CHAP TER 1 INTRODUCTION .................................................................................................................. 19 2 THE EFFECT OF PROTEIN BINDING ON THE PHARMACOLOGICAL ACTIVITY OF HIGHLY BOUND ANTIBIOTI CS .............................................................. 23 Introduction .................................................................................................................. ...........23 Material and Methods .............................................................................................................24 Organisms ..................................................................................................................... ...24 Antibiotics and Growth Media ........................................................................................ 25 HPLC Analysis ................................................................................................................ 25 Instrumentation ......................................................................................................... 25 Chromatographic conditions ....................................................................................26 Protein Binding ................................................................................................................26 Pharmacodynamics .......................................................................................................... 27 MIC .......................................................................................................................... 27 Time-kill cu rves ....................................................................................................... 28 Mathematical Modeling ...................................................................................................28 Results .....................................................................................................................................29 Protein Binding ................................................................................................................29 Pharmacodynamics .......................................................................................................... 29 MIC .......................................................................................................................... 29 Time-kill cu rves ....................................................................................................... 30 Mathematical Modeling ...................................................................................................30 Discussion .................................................................................................................... ...........30 3 CLINICAL MICRODIALYSIS IN SKIN AND SOFT TISSUES: AN UPDATE ................42 Introduction .................................................................................................................. ...........42 Rationale for this Review .......................................................................................................43 Methods ..................................................................................................................................44
7 Calibration Methods ........................................................................................................ 44 Strengths of the Microdialysis Technique ....................................................................... 45 Limitations of the Microdialysis Technique .................................................................... 46 Bioavailability at the Site of Action ................................................................................47 Factors Affecting Bioavailability .................................................................................... 53 Bioequivalence ................................................................................................................ 55 PK/PD Indices .................................................................................................................56 T>MIC .......................................................................................................................58 AUC24/MIC ............................................................................................................ 59 Cmax/MIC ................................................................................................................ 60 Conclusion .................................................................................................................... ..........62 4 PENETRATION OF ERTAPENEM INTO SKELETAL MUSCLE AND SUBC UTANEOUS ADIPOSE TISSUE IN HEALTHY VOLUNTEERS MEASURED BY IN VIVO MICRODIALYSIS ...........................................................................................70 Introduction .................................................................................................................. ...........70 Volunteers and Methods ........................................................................................................ .71 Volunteers .................................................................................................................... ....71 Study Design and Protocol ..............................................................................................72 Sample Collection ...........................................................................................................72 Drug Assay ......................................................................................................................74 Pharmacokinetic Analysis ...............................................................................................74 Results .....................................................................................................................................75 Safety ...............................................................................................................................75 Pharmacokinetics ............................................................................................................. 76 Discussion .................................................................................................................... ...........76 5 INTEGRATION OF MODELING AND SI MULATION IN DEVEL OPMENT OF NEW ANTI-INFECTIVE AGENTS MI C VS. TIME-KILL CURVES ............................. 82 Introduction .................................................................................................................. ...........82 PK/PD Strategies for Anti-Infectives .....................................................................................83 Effect of Drug Binding .................................................................................................... 83 MIC-Based PK/PD Indices .............................................................................................. 84 Time-Kill Curve-Based PK/PD Indices .......................................................................... 85 Pharmacodynamic Models .............................................................................................. 87 PK/PD Simulations .......................................................................................................... 91 Monte Carlo simulation ............................................................................................91 Time-kill curve based simulations ........................................................................... 93 Discussion and Conclusion .....................................................................................................94
8 6 PHARMACOKINETIC/PHARMACODYNAMIC MODELING OF THE IN VITRO ACTIVITY OF OXAZ OLIDINONE ANT IBIOTICS AGAINST METHICILLINRESISTANT STAPHYLOCOCCUS AUREUS .....................................................................102 Introduction .................................................................................................................. .........102 Material and Methods ...........................................................................................................103 Antibiotics and Growth Media ...................................................................................... 103 Organisms ..................................................................................................................... .104 MIC Determination .......................................................................................................104 Constant Concentration Time-Kill Curves .................................................................... 104 Changing Concentration Time-Kill Curves ................................................................... 105 Drug Stability ................................................................................................................106 Mathematical Modeling .................................................................................................106 Data Analysis .................................................................................................................107 Model Validation ...........................................................................................................108 Results ...................................................................................................................................108 MIC ........................................................................................................................... .....108 Constant Concentration Time-Kill Curves .................................................................... 108 Changing Concentration Time-Kill Curves ................................................................... 109 Drug Stability ................................................................................................................109 Mathematical Modeling .................................................................................................109 Model Validation ...........................................................................................................109 Discussion .................................................................................................................... .........110 7 DISCUSSION .................................................................................................................... ...118 LIST OF REFERENCES .............................................................................................................123 BIOGRAPHICAL SKETCH .......................................................................................................147
9 LIST OF TABLES Table page 2-1 Determined MIC values (presented as m odes) and simultaneously fitted model parameters (SD) of ceftriaxone and ertapenem against E. coli and S. pneum. in the presence and absence of bovine serum albumin (BSA) and human serum albumin (HSA), respectively. ...........................................................................................................36 3-1 Pharmacokinetic parameters for diclofenac in plasm a and subcutaneous and skeletal muscle tissue. .....................................................................................................................64 4-1 Non-compartmental pharmacokinetic anal ysis of ertapenem after 1 g single intravenous dose............................................................................................................... ..80 5-1 Pharmacokinetic and pharmacodynamic pa ra meters for ceftriaxone and faropenem used for the simulations. .................................................................................................... 97 6-1 Comparison of the final model paramete r estimates (MSE) and estim ates (95%CI) from 1000 nonparametric bootstrap runs ......................................................................... 114
10 LIST OF FIGURES Figure page 2-1 In vitro protein binding of ceftr iaxone and ertapenem ...................................................... 38 2-2 Effect of bovine (BSA) and hum an serum albumin (HSA) on bacterial growth and antibiotic-induced kill ........................................................................................................39 2-3 Simultaneous curve fits for ceftriaxone against E. coli and S. pneum. in the presence and absence of bovine serum albu min and human serum albumin ................................... 40 2-4 Simultaneous curve fits for ertapenem against E. coli and S. pneum. in the presence and absence of bovine serum albu min and human serum albumin ................................... 41 3-1 Mean concentration-tim e profile of diclofenac after oral an d topical administration in subcutaneous tissue and plasma ......................................................................................... 65 3-2 Concentration-tim e plots demonstrating differences in dermal salicylic acid penetration sampled by MD probes, inserted in the 4 barrier-pertubated skin areas ......... 66 3-3 Median con centration-time profiles for penciclovir in skin for control, solution perfused with adrenaline, and cold skin following single oral administration of 400mg famciclovir (prodrug) in healthy volunteers ..........................................................67 3-4 Ertapenem concentration-time profiles in to tal plasma, skeletal muscle fluid, and interstitial adipose tissue following 1g infu sion for 30min in healthy volunteers ............. 68 3-5 Telith romycin concentration-time profiles in plasma, muscle, and subcutaneous adipose tissue after a single 800mg dose in healthy volunteers ......................................... 69 4-1 Comparison of ertapenem concentration profiles in plasma with unbound tissue concentrations in skeletal muscle fluid and interstitial adipose tissue fluid in healthy volunteers after a single intravenous dose of 1 g ............................................................... 81 5-1 Design of in vitro m odel .................................................................................................... 98 5-2 Two-com partment body model with additional effect compartment ................................ 99 5-3 Growth/kill curves a t different dr ug concentrations to illustrate k0, kmax, EC50, MIC and SC relationship ..........................................................................................................100 5-4 Simulated time-kill curves and respective growth controls of faropenem against Haemophilus influenzae ATCC10211 and ceftriaxone against Streptococcus pneumoniae CDC145 .......................................................................................................101 6-1 Susceptibility-based two-subpopulation m odel ............................................................... 115
11 6-2 Simultaneous curve fits of the susceptibility-based two-compartment model to the experimental data ............................................................................................................. 116 6-3 Basic diagnostic plots ......................................................................................................117 7-1 Interplay between pharmac okinetics, pharm acodynamics, patient and disease .............. 122
12 LIST OF ABBREVIATIONS ABS Adult bovine serum AIC Aminoimidazole-4-carboxamide ALA Aminolevulinic acid ANOVA Analysis of variance AUC Area under the curve AUC0-t Area under the curve from time zero to time t AUC0Area under the curve from time zero to infinity AUC/MIC Area under the curve over 24h at steady-state divided the MIC AUMC0Area under the first-moment curve BA Bioavailability BAL Broncho-alveolar lavage BCC Basal cell carcinoma BE Bioequivalence BID Twice daily BSA Bovine serum albumin CE Conformite Europeene CDC Center for Disease Control and Prevention CFU Colony forming units CI Confidence interval CLSI Clinical and Laboratory Standards Institute CLtot Total clearance Cmax Peak plasma concentration
13 Cmax/MIC Peak level divided by the MIC COX-2 Cyclooxygenase 2 CRO Ceftriaxone dg Delay in the onset of growth dgs Delay in the onset of grow th of susceptible bacteria dk Delay in the onset of kill dks Delay in the onset of ki ll of susceptible bacteria E. coli Escherichia coli EC50 Concentration necessary to produce 50% of the maximum (kill) effect ECG Electrocardiogram EE Extraction efficiency ELF Epithelial lining fluid Emax Maximum effect ERT Ertapenem ESBL Extended-spectrum -lactamase Free Ce Free concentration in the effect compartment FDA Food and drug administration fu Fraction unbound GI Gastro intestinal h Hill factor H. influenzae Haemophilus influenzae HSA Human serum albumin
14 HP Human plasma HPLC High performance liquid chromatography ISF Interstitial space fluid IRB Institutional review board i.v. Intravenous infusion k0 Bacterial growth-rate constant k12 Transfer-rate constant from the first to the second compartment k21 Transfer-rate constant from the second to the first compartment k10 Elimination-rate constant from the first compartment k1e Transfer-rate constant from the first in the effect compartment ka Absorption-rate constant kd Natural death-rate constant ke0 Elimination-rate constant from the effect compartment K. pneumoniae Klebsiella pneumoniae kmax Maximum kill effect kps Transfer-rate constant from persistent stage to susceptible stage ks Growth-rate constant of susceptible pathogens ksp Transfer-rate constant from sus ceptible stage to persistent stage LC-MS-MS Liquid chromatography coupled with tandem mass spectrometry LR Lactated ringers z Terminal elimination rate constant MHB Mueller-Hinton broth MIC Minimum inhibitory concentration
15 MD Microdialysis MMM Metastatic malignant melanomas MPT Mechanical pain threshold MRSA Methicillin-resistant Staphylococcus aureus MRT Mean residence time MSC Model selection criterion MSE Mean standard error N Number of bacteria n Number of study subjects N0 Starting number of bacteria NCCLS National Committee for Clinical Laboratory Standards NMIC Number of bacteria at the MIC turbidity threshold Np Number of bacteria unsusceptible to antibiotic Ns Number of bacteria su sceptible to antibiotic NSAID Nonsteroidal anti-inflammatory drug Nt Number of bacteria at time t OFV Objective function value P. aeruginosa Pseudomonas aeruginosa PB Protein binding PCV Penciclovir PDT Photodynamic therapy PG Prostaglandin PET Positron emission tomography
16 PD Pharmacodynamic PK Pharmacokinetic QD Once daily S. aureus Staphylococcus aureus SC Stationary concentration SD Standard deviation SLS Sodium lauryl sulfate S. pneum. Streptococcus pneumoniae SSSI Skin and skin structure infections t Time T>MIC Cumulative percentage of a 24h period that the drug level exceeds the MIC at steady-state t1/2 Half-life TB Tuberculosis THB Todd-Hewitt broth TID Three times daily tmax Time to reach peak plasma concentration Vd Volume of distribution VISA Vancomycin-intermediate susceptible Staphylococcus aureus VRE Vancomycin-resistant enterococci VRSA Vancomycin-resistant Staphylococcus aureus Vz Apparent volume of distribution during the terminal phase
17 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PHARMACOKINETICS AND PHARMACO DYNAMICS OF OXAZOLIDINONES AND BETA-LACTAMS By Stephan Schmidt December 2008 Chair: Hartmut Derendorf Major: Pharmaceutical Sciences Introduction: The purpose of this thesis was to s upport optimizing antibiotic drug development by evaluating the effect of pr otein binding (PB) on 1) the in vitro activity of highly bound agents and 2) on the soft tissue penetra tion, as well as, 3) establishment of a general PK/PD model to characterize the activity of oxazolidi nones against methicillin-resistant Staphylococcus aureus (MRSA). Methods: 1) Using in vitro microdialysis (MD), free ceftria xone (CRO) and ertapenem (ERT) concentrations were determined in growth medium with and without bovine/human serum albumin (BSA/HSA), in adult bovine serum (A BS) and in human plasma (HP). Corresponding antimicrobial activity was determined in minimum inhibitory concentration (MIC) and time-kill curve experiments against E. coli ATCC25922 and S. pneum. ATCC6303. A modified Emaxmodel was fitted to the data and respective EC50s compared. 2) Blood and MD samples from the interstitial space fluid (ISF) of thigh muscle a nd subcutaneous adipose tissue were obtained from 6 healthy volunteers following a 30min infusi on of 1g ERT for up to 12 hours. 3) A twosubpopulation model was simultaneously fitted to th e static, as well as, dynamic time-kill curve data of RWJ-416457 and linezolid against MRSA OC2878 and EC50s compared.
18 Results: 1) For CRO, PB differed between HP (76.8.0%) and commercially available BSA (20.2.3%) or HSA (56.9.6%). Similar resu lts were obtained for ERT (HP: 73.8.6%, BSA: 12.4.8%, HSA: 17.8.5%). MICs and EC50s of both strains were increased for CRO when comparing HSA and BSA, whereas EC50s were not different for ERT. 2) Free concentrations in the ISF of thigh muscle and su bcutaneous adipose tissue were consistent with free concentrations in plasma but much lowe r than those in total plasma. 3) A twosubpopulation model was appropriate to desc ribe the data resulting in a lower EC50 of RWJ416457 (0.41 g/mL) compared to that of linezolid (1.39 g/mL). Conclusion: Free, unbound concentrations are responsible for antimicrobial efficacy, as well as, tissue distribution and can be measured by microdi alysis. Once determined, free concentrations can be linked to the corresponding PD parameters in an appropriate PK/PD model in order to predict clinical outcome. PK/P D modeling and simulation is, cons equently, a valuable tool for dose selection.
19 CHAPTER 1 INTRODUCTION An infection is defined as the growth of a parasitic organism (e.g. bacteria, fungi, etc.) within the body of a host. The goal of an anti-infec tive therapy is to er adicate the infecting organism (pathogen) from the host. Generally, this removal process primarily includes the use of anti-infective drugs. Ideally, th ese drugs eliminate the pathog en without causing toxic side effects or inducing resistance against the anti-inf ective agent in the pathog en. The selection of an appropriate antimicrobial agent and/or dosing regimen is hereby crucial for success of the therapy and can be governed by pharmacokinetic (PK) and pharmacodynamic (PD) principles. Hypothesis: Insufficient free, unbound antibiotic concen trations at the si te of infection result in therapeutic failure a nd can, subsequently, lead to the emergence of resistance. Proper selection of an appropriate antibiotic dosing regimen is, therefore, critically important and can be supported by pharmacokinetic/pharmacodynamic (P K/PD) modeling and simulation approaches. The selection of appropriate antimicrobial agents does not start at the bedside but way early in drug development by evaluating the antim icrobial spectrum of ne w antibiotic candidates in the laboratory. This is usually done by dete rmining the minimum inhibitory concentration (MIC) against a variety of Gram-positive, Gram -negative and atypical pathogens. The MIC is defined as the lowest concentration that complete ly inhibits visible growth of the organism, as detected by the unaided eye after overnight incubation at 35oC with a standard inoculum of approximately 5x105-106 colony forming units per milliliter (CFU/mL) (228). Although routinely employed, the MIC is a static, mono-dimens ional threshold value that does not allow to account for changes of antibiotic co ncentrations over time or pred iction of antimicrobial activity at concentrations apart from the MIC itself (148, 216, 231). More detailed information can be obtained from evaluation of the antimicrobial activity over time (t ime-kill curves).
20 In order to facilitate the transition between id entification of antibiotic candidates and later, clinical stages of drug development, employed in vitro systems should be as predictive as possible of clinical outcome. One appr oach is to reproduce similar conditions in vitro as encountered in vivo For example, addition of protein supplements to the in vitro MIC or timekill curve test systems is intended to mimic plasma protein bindi ng. However, the experimental procedure for the determination of protein binding has not been internationally standardized yet and the use of different supplements might yield different outcomes. The first specific goal of this thesis work is the evaluation of the antimicrobial activity of two beta-lactams with reportedly high plas ma protein binding, ceftriaxone (83-96%) (244, 245, 277) and ertapenem (84-96%) (34), in the pr esence of various protein supplements. Binding to plasma proteins does not only reduce the antimicrobial activity but also restricts the distribution of antibiotics into tissues. Since most infections are located in the tissues rather than the blood stream, sufficient penetration into the tissue is essential for an antibiotic to be clinical effective and, subsequent ly, to avoid the emergence of re sistance (251). The traditional approach of linking plasma drug co ncentrations to observed effects is, consequently, unreliable (185, 251). A more rational approa ch would be to determine free, active drug concentrations at the site of infection. Microdialysis (MD) is currently the most appropriate sampling technique that can provide this information and has been frequently employed in PK studies determining the tissue concentrations of antibiotics (44, 112, 1 20). In this sampling technique, a small, semipermeable membrane (MD probe) is placed into th e ISF of the tissue of interest. The MD probe is perfused with a physiological solution (e.g. Lact ated Ringers) at a co nstant flow-rate of 15 L/min and at specified time intervals protein-fr ee compound is collected for analysis (28, 228).
21 The second specific goal of this thesis work is to determine the dist ribution of ertapenem into muscleand subcutaneous adipose fat tissue of 6 healthy subjects fo llowing a single 30min infusion of 1g ertapenem using microdialysis. At this point it is important to realize that in isolation, PK and PD information is of limited usefulness. Only the link of PK and PD allows to sufficiently characterizing and predicting the drug-effect relationship. For antibiotics, a co mbination of the MIC and free (), unbound PK parameters to MIC-based PK/PD indi ces is frequently used to defi ne clinically relevant efficacy breakpoints. To date, three main PK/PD indices are employed and have led to a much better understanding of antibiotic dosing: 1) the time free antibiotic concentrations remains above the MIC (T>MIC), free area under the concen tration-time curve over MIC ratio (AUC/MIC) and free maximum plasma concentration over MIC ratio (Cmax/MIC) (229). Yet, the previously described drawbacks of MIC (static, highly variab le, etc.) (231), limit the reliability of the MICbased PK/PD indices as well. To overcome these limitations, the use of ot her PD approaches has been suggested. One of these approaches is the continuous measurement of the antibiotic concentration-effect relationship over time ( time-kill curves) (11, 198, 258). Once performed, a mathematical model can be simultaneously fit to the data and outcome parameters, such as the EC50 (concentrations necessary to produce 50% of the maximum effect) can be used to characterize the potency of antibiotics. Although the description of these time-kill curves is mathematically somewhat more complex, their capability of describing concentration-effect relationships over time allow the evaluation of su b-MIC, as well as, fluc tuating concentrations on bacterial growth and kill. Once a mathematical model is established, it can be applied to the time-kill curve data of other drugs with the same mechanism of action and respective PD outcomes can be compared. In addition, the link of determined PD parameters to PK data of
22 different doses or dosing regimens allows predicting corresponding clinical outcomes. Modeling and simulation approaches are, consequently, ve ry valuable for dose selection and have been recommended by the Food and Drug Administrati on (FDA) as tools to streamline the drug development process (84). The third specific goal of this thesis work is therefore, to qualita tively and quantitatively compare the antimicrobial activity of RWJ-416457 to that of the first-in-class representative linezolid. For qualitative comparison, MIC and constant, as well as ch anging concentration time-kill curve experiments of linezolid and RWJ-416457 against MRSA OC2878 will be performed. For quantitative comparison, a genera l PK/PD model will be established, validated and simultaneously fitted to the tim e-kill curve data and respective EC50 values compared.
23 CHAPTER 2 THE EFFECT OF PROTEIN BINDING ON THE PHARMACOLOGICAL ACTIVITY OF HIGHLY B OUND ANTIBIOTICS1 Introduction Most drugs bind to protei ns or other biological m a terials such as albumin, 1-acid glycoprotein, lipoproteins, -, and globulins and erythrocytes. Thus, free, unbound drug concentration in plasma decrease s as the degree of binding to thes e compounds increases. It is a well recognized fact that, at le ast for small molecules, only free, unbound drug distributes into the extravascular space and is responsible for ph armacological activity and/or side effects (200, 224, 280). For antibiotics in particular, reduced fr ee, active drug concentra tions as a result of protein binding reflect in decrease d antimicrobial activity (200). In theory, this effect is most pronounced for antibiotics with ex tensive protein binding. To date, most studies evaluating the effect of protein binding on the potency of an antibiotic against a certain pathogen determine changes in the respective minimum inhibitory concentration (MIC) (200). Whereas both bacterial growth and kill are dynamic processes, the MIC is a static, high ly variable threshold value, incapable of predicting an tibiotic activity at c oncentrations apart from the MIC (183, 229). In comparison, evaluation of growth and antib iotic-induced kill profiles over time (time-kill curves) provides more detailed information th an the MIC. Once experimentally determined, time-kill curves can be characterized by simultaneous fit of appropriate mathematical models and quantitatively compared by respective outcome parameters, such as the EC50 (229). Time-kill curves have, consequently, been suggested as an experimental method for the evaluation of protein binding effects on the an timicrobial activity (231, 280). In order to account for protein binding in these experiments, bacterial media ar e spiked with either human serum or protein 1 Copyright American Society for Microbiology, [Antimicrob Agents Chemother. 52: 3994-4000, 2008]
24 supplements. When supplementing with human seru m, its actual content frequently has to be limited to 50% since it may inhibit bacterial growth or modify the antibacterial activity (140, 144, 200). On the other hand, when supplementing with proteins, usually human serum albumin (HSA) or comparatively less expensive animal albumins (20, 101, 111) are employed, as HSA is the main natural binding component of antib iotics (142, 218, 247, 262, 280). Nevertheless, the actual free, unbound antibiotic concen tration after either HS or pr otein supplementation is rarely determined. Instead, literature or protein binding values determined in vitro are frequently employed to estimate free, unbound concentrations (36, 141). The goal of the present study was to evaluate this approach by linking measured free, unbound concentrations of the two highly-bound -lactams ceftriaxone (83-96%) (244, 245, 277) and ertapenem (84-96%) (34) to their respectiv e antimicrobial activity against Gram-positive Streptococcus pneumoniae ( S. pneum.) ATCC6303 and Gram-negative Escherichia coli ( E. coli. ) ATCC25922. Material and Methods Organisms E. coli ATCC25922 and penicillin-sensitiv e S. pneum. ATCC6303 were obtained from the Clinical Microbiology Laboratory at Shands Hospital at the Un iversity of Florida, USA. E. coli and S. pneum. were grown in a CO2 incubator (Barnstead-Thermol yne, Melrose Park, IL, USA) in Mueller-Hinton broth (MHB) or Todd-Hewitt broth (THB) plus 5% CO2, respectively. To ensure purity of the bacterial strains, they were subcultured at least three times before either usage or freezing of the stock cultures. The b acterial inoculum was prepared from colonies incubated overnight on 5% sheep blood agar plat es (Remel Microbiology Products, Lenexa, KS, USA). The microorganisms were suspended in st erile saline solution 0.9% to a concentration equivalent to a 0.5 value in the McFarland scale (Remel Microbiology Products, Lenexa, KS,
25 USA) with a turbidimeter (A-JUSTTM, Abbott Laboratories, North Chicago, IL, USA). This value on the McFarland scale of 0.5 is equivalent to a number of 1x108 viable colony forming units per milliliter (CFU/mL). Further dilution steps to reach a final working inoculum of approximately 5x105 CFU/mL were performed in broth. Antibiotics and Growth Media Ceftriaxone disodium was purchased from Sigma-Aldrich (St. Louis, MO, USA) and ertapenem sodium was obtained from Merck & Co., Inc. (Whitehouse Station, NJ, USA). Antibiotics were prepared and stored accordi ng to the manufacturers recommendations. MHB (Becton Dickinson, Franklin Lakes, NJ, USA) and THB (Difco, Detroit, MI, USA) were used as liquid growth media. MHB a nd THB were both prepared acco rding to the manufacturers instructions and autoclav ed prior to use at 121oC (15 min per 1L). Broth media were supplemented with either 40g/L BSA (Sigma-Ald rich, St. Louis, MO, USA, Cat. No. A3059, Lot. No. 036K0735) or 40g/L HSA (Calbiochem, La Jolla, CA, USA, Cat. No. 12666, Lot No. B75308-01), filtered through 0.2 m filters (Millipore, Billerica, MA, USA) and the pH adjusted to 7.4. HPLC Analysis Instrumentation The HPLC-s ystem for both ceftriaxone and er tapenem consisted of Agilent 1100 Series (Agilent Technologies, Waldbronn, Germany): a model G1313 autosampler, a model G1311 quaternary pump and G1315 DAD UV detector, an Agilent Chemstation for LC systems and a LiChrospher 100 Reversed Phase 18 (RP-18, 5m partic le size) analytical column (Merck KGaA, Darmstadt, Germany).
26 Chromatographic conditions Ceftriaxone: The ion-pair chrom atography a ssay procedure was adapted from Kovar et al..(133) Reversed phase high-performance liqui d chromatography (RP-HPLC) was performed at room temperature at a flow-rate of 1.0mL/min. The mobile phase consisted of a mixture of buffer (7.5mM KH2PO4; J.T. Baker Chemical Co., Phillipsburg, N.Y., USA) and acetonitrile (56:44, v/v) with 5mM hexadecyltrimethyla mmonium bromide (HDTA; Sigma-Aldrich, St. Louis, MO, USA) as the ion-pair reagent. The final pH of the mobile phase was adjusted to 8.8. 25 L of ceftriaxone were injected and detected at 280nm. The run time was set to 16 minutes. Ertapenem: The chromatography procedure was adapted from Gordien et al. (98). The mobile phase consisted of 10mM phosphate bu ffer adjusted to pH 6.5 with concentrated orthophosphoric acid and mixed with acetonitrile. A gradient was run at a flow-rate of 1mL/min. 40 L of ertapenem were injected and detected at 305nm. The run time was set to 14 minutes. Protein Binding Free, protein-unbound ceftriaxone and ertapenem concentrations were determ ined in in vitro dose-ranging (20, 40, 80, 160 and 320 g/mL) extraction efficiency (EE) microdialysis experiments. Briefly, blank Lactated Ringer (L R; Baxter Health Care Deerfield, IL, USA) solution is pumped through a flexible CMA 60 microdialysis probe (CMA Microdialysis AB, Solna, Sweden) at a constant flow rate of 2.0 L/min. The relative recovery (RR) of the microdialysis probes were determined in LR only and calculated according to equation 2-1, sample dialysateC(LR) C(LR) RR (2-1) where C(LR)dialysate is the free concentration recovere d from the microdialysis probe and C(LR)sample the concentration in the test tube.
27 Once RR was determined, free, unbound ceftriaxone and ertapenem concentrations were determined in triplicate at 37oC in THB, THB with BSA, THB with HSA and pooled adult bovine serum (ABS; HyClone, Logan, Utah, USA) or pooled human plasma (HP; Shands Hospital at the University of Florida, Gainesville, USA), respectively. It was previously reported that serum binding properties can be altered by heat treatment (269). Therefore, untreated pooled HP was used for the in vitro protein bindi ng experiments. Protein binding (%) of the respective samples can then be calculated according to equation 2-2, total dialysateC(sample) RR C(sample) 1PB(%) (2-2) where C(sample)dialysate is the concentration in the pr otein-free dialysate and C(sample)total is the mean concentration of the samples that were collected directly out of the test tube at the beginning and the end of the 30min sampling period. Collected samples were analyzed by the RP-HPLC methods described above. Protein binding differences within the treatment groups were evaluated using analysis of variance (ANOVA) followed by least square means to test for pair-wise differences. All statistical analysis was performed in SAS 9.1.3 (SAS Institute Inc., Cary, USA). A P-value <0.05 was considered statistically significant. Pharmacodynamics MIC MICs of E. coli ATCC25922 and S. pneum. ATCC6303 against ceftriaxone and ertapenem were determined six times according to the CSLI guidelines both in presence and absence of BSA and HSA using a serial dilution twofold Macro Broth Dilution method (258).
28 Time-kill curves In vitro 6 hour constant concentration time-kill curves were performed in triplicate for both ceftriaxone and ertapenem against the test strain s in the presence and absence of BSA and HSA, respectively (258). Eight 50 mL cell culture flasks (NuncTM, Nunc A/S, Roskilde, Denmark) were filled with 20mL of bateria-containing growth medium and incubated for two hours before adding the antibiotic. The selection of respective ceftriaxone or ertapenem concentrations (Table 3-1) was based on their determined MIC values and covered the entire antimicrobial spectrum, including minimum inhibition of bacterial grow th (0.25, 0.5, and 1 times MIC), efficient bacterial killing (2 and 4 times MIC) and maximum bacterial killing (8 and 16 times MIC). Samples were taken at 0, 0.5, 1, 1.5, 2, 3, 4, 5 and 6 hours. Bacteria l counts were determined, using an adapted droplet-plate method (258). A control experiment with bacteria and no drug was run simultaneously. After incubation at 37oC for 20-24hours, viable counts were determined on all readable plates. Mathematical Modeling A modified susceptibility-based two-compar tment model was simultaneously fit to the time-kill curve data of with and without BSA and HSA, respectively (198, 258). In this model, the overall change in the experimentally determined total number of bacteria was defined as the sum of self-replicating, antibiotic-sensitive ce lls and metabolically inactive, insusceptible persister cells (10). While bacterial growth c ould be sufficiently described by the growth-rate constant ks (h-1), antibiotic-induced kill was characteri zed by the maximum kill-rate constant kmax (h-1), the antibiotic concentration C ( g/mL) and the concentration at half-maximum effect EC50 ( g/mL). The final shape of the curve fit could be optimized by a Hill or shape factor h. In order to compare the respective EC50 values in the presence or absence of BSA or HSA, ks, kmax and h values were fitted across all treatment groups for each antibiotic and strain. Initial
29 parameter estimates for ks, kmax and h were obtained from individual curve fits and compared for differences. Since no differences in paramete r estimates were observed (data not shown), arithmetic means were used as the initial estim ates for the simultaneous curve fits across the treatment groups. After simultaneously fitting the susceptibility-based sigmoidal Emax-model to the experimental time-kill curve data with the non-l inear least-square regression software Scientist 3.0 (Micromath, Salt Lake City, UT, USA), m odels were characterized by Model Selection Criteria (MSC) and graphs visually inspected for quality of fit. EC50 comparisons were done using ANOVA, followed by least square means. A P-value <0.05 was considered statistically significant. Results Protein Binding As shown in Figure 2-1, mean protein binding (SD) values of both ceftriaxone (76.8.0%) and ertapenem (73.8.6%) were hi gher in pooled HP than those in THB with and without BSA or HSA, as well as, pooled ABS, respectivel y (P<0.05). For ceftriaxone, protein binding was significantly lower in co mmercially available BSA (20.2.3%) and HSA (56.9.6%), as well as, pooled ABS (30.7.2%). Similar lower protein binding values were also observed for ertapenem (BSA: 12.4.8%, HSA: 17.8.5%, pooled ABS: 38.3.8%). Pharmacodynamics MIC Determined MIC values (presented as modes) are shown in Table 2-1. No differences in MIC values were found for ertapenem against S. pneum. both in the presence and absence of BSA and HSA, respectively. However, in the presence of HSA (vs. no albumin), MICs were increased for ceftriaxone against E. coli and S. pneum., as well as, for ertapenem against E. coli..
30 In comparison, when supplementing with BSA (vs. no albumin), no change in MICs was detected for ceftriaxone against E. coli and ertapenem against E. coli, whereas values were increased for ceftriaxone against S. pneum.. Time-kill curves Qualitative evaluation of the time-kill curves showed that, in comparison to albumin-free medium, there are no differences in growth of both E. coli and S. pneum. in the presence of BSA and HSA (Figure 2-2), respectivel y. However, further evaluati on showed differences in antimicrobial activity when comparing samples with and without albumin supplementation. These differences were most apparent at concentrations of 8 times MI C (with respect to albumin-free medium). While the maximum kill effect was not reached at 8xMIC for ceftriaxone when HSA was added, these concentrations were sufficient for ertapenem and ceftriaxone in the presence of BSA as well as for ertapenem supplementing with HSA. Mathematical Modeling Simultaneous curve fits of ceftr iaxone and ertapenem against E. coli and S. pneum. with and without BSA or HSA are shown in Fi gure 2-3 and 2-4. The corresponding model parameters for ceftriaxone and ertapenem agains t both strains in the presence and absence of BSA and HSA are listed in Table 21. For ceftriaxone, calculated EC50 values for both strains were higher in the presence of HSA than those wi th BSA (P<0.05) or without albumin (P<0.05), respectively. The difference remained significan t after adjusting for multiple comparisons. In contrast, no differences in EC50 values were determined for ertapenem against both strains when it was supplemented with HSA compared to no albumin and BSA. Discussion The clinical significance of protein binding on antimicrobial activity continues to be controversial due to conflicting reports from in vitro MIC and/or time-kill curve experiments
31 (232). In these experiments, bact erial media are spiked with eith er serum or protein supplements in order to produce and modify protein binding. However, whet her the degree of drug binding in these in vitro test systems is representative of the respective physiological conditions is often unclear since free, unbound concentrations are fr equently not experimentally determined. Instead, reported literature values are commonly used to correct to tal concentrations for protein binding (36, 141). The results of our study, howev er, clearly show that this approach can be extremely misleading. In this study, MICs and constant concentra tion time-kill curves of two beta-lactams with reportedly high protein binding, ceftriaxone (83-96%) (244, 245, 277) and ertapenem (84-96%) (34), were determined in the presence and ab sence of BSA and HSA (40g/L), respectively, and outcomes compared. Results indicate that th e antimicrobial activity of both ceftriaxone and ertapenem was decreased against both test strains, except for ertapenem against S. pneum., in the presence of HSA. However, MICs remain ed unchanged, except for ceftriaxone against S. pneum., in the presence of BSA. In theory, these differences in pharmacodynamic outcome could be explained by albumin-i nduced effects on growthand ma ximum kill-rates, as well as, changes in potency (EC50) (188, 200, 231). The evaluation of bacterial growth is, thereby, of particular importance since only dividing cells are susceptible to beta-lactams (200). However, evaluation of the respective growth controls ov er time revealed that there were no significant differences in growth-rates of both test strain s in the presence or absence of BSA and HSA, respectively (Figure 2-2). Once antibiotic is added, antimicrobial agents ki ll bacteria more rapidly as concentrations increase.(211) Yet, at concentr ations of 2-4 times the MIC, the respective response varies (211). For beta-lactams, the maximum kill effect is already observed. At this point, a further increase
32 in antibiotic concentrations does not result in incr eased killing and the kill rate remains constant. The findings of our study are in agreement with these concepts. However, results further revealed that ceftriaxone concentrations of 8 tim es MIC (with respect to no albumin) are not high enough to reach kmax when HSA is added, but are sufficien t when BSA is present (Figure 2-2). When further increasing concentrations to 16-32 times MIC, the maximum kill effect was reached and no albumin-related effects on kmax were determined. In contrast, saturation in kill was achieved for ceftriaxone and ertapenem in th e presence of BSA as well as for ertapenem when adding HSA at 8xMIC. Nevertheless, substantial differences in mean EC50s were determined for ceftriaxone, whereas, no significant differences in mean EC50s were determined for ertapenem in the presence of HSA. In comparison, no differences in mean EC50 values were determ ined for both strains and antibiotics when it was supplemented with BSA. At this point, it is im portant to realize that in isolation, the results of the MIC and time-k ill curves lead to ambiguous conclusions about the impact of protein binding on the antimicrobial ac tivity. While a two-fold increase in MIC was observed for both ceftriaxone against S. pneum. in the presence of BSA and ertapenem against E. coli in the presence of HSA, no significant differences in EC50s were determined in the respective time-kill curve experiments. These di fferences in pharmacological outcome (MIC vs. EC50) are frequently attributed to the immanent high variability of the employed Macro-broth dilution (two-fold) method (188, 229). Nevertheless, the fact that there are tremendous differences in both MICs and EC50s between ceftriaxone and ertapenem against both strains cannot simply be explained by variability or calcul ated free concentrations since both antibiotics have very similar reported protein binding (ceftriaxone: 83-96%, ertapenem: 84-96%) values (34, 244, 245, 277). However, the major assumption that the reported in vivo protein binding
33 values reflect also the binding conditions in the in vitro system is rarely validated by experimentally measuring free, unbound concentrations. Different methods, such as, equilibrium dialys is, ultrafiltration, micr odialysis, etc. have been used for the determination of protein bi nding and have shown comparable outcomes (12, 103). In our study free, unbound concentrations were measured by in vitro microdialysis and respective protein binding values were calculat ed (Figure 2-1). Results indicate that the observed differences in antimicrobial activity of ceftriaxone and ertape nem can be explained by differences in their in vitro binding to the respective albumin supplements. While ceftriaxone is extensively bound to the tested HSA, it shows onl y little binding to BSA, and ertapenem hardly binds to either one. The findings of our in vitro microdialysis study are in agreement with those from Nix et al., where the bi nding of ertapenem to purified albumins was determined by ultrafiltration (200). Findings of both studies concurrently indicate that the in vitro binding to albumin supplements was substantially lower than previously reported literature protein binding values. Results of the ultrafiltration experiment further showed that th e binding of ertapenem to various albumin supplements was greatly depe ndent on the albumin preparations used and differed substantially between suppliers (200). These observations may be explained by the lack of fatty acids in the albumin supplements (246) and/or the use of rigor ous conditions, such as, heat treatment or organic solvents during th e purification process that can result in conformational changes in the respective albumin binding sites and, subsequently, reduced protein binding values (200). On the other hand the significantly lower binding to pooled ABS, compared to pooled HP, indicates that for bot h antibiotics differences in binding capacities between species may play a role. For ceftria xone, similar protein bindi ng values have been previously determined in human, rat, baboon and rabbit plasma, whereas substantially lower
34 binding values were measured in dog plasma (212). This article by Popick et al. further revealed that the initially high protei n binding of ceftriaxone (9095%) at concentrations <100 g/mL is considerably decreased to approximate ly 60% at higher co ncentrations (>400 g/mL) (212). In our study, mean protein bindi ng values (ceftriaxone: 76.8 11.0%) were determined at concentrations ranging from 20-320 g/mL and are in agreement with the previously reported range. At this point, it also should be mentioned that there are other factors that may provide an explanation for the differences in reported outc omes that range from no protein binding effects (88, 141), delay in the onset of activity (36, 41), to only free, unbound drug being responsible for the antimicrobial activity (177, 200, 224, 261, 279, 280). For example, it has been shown that the bacterial density or the state of nutrition are crucial parameters and us ually differ in MIC and time-kill curve studies from the physiological co nditions. While a lack of nutrition and low bacterial numbers hinder the experimental conduct, ve ry high initial inoculum sizes seem to alter the susceptibility towards antimic robial agents and might mask pr otein binding effects (169). It would, therefore, be importan t to internationally standardi ze the methodology of protein binding studies in order to minimize the experimental bias. In conclusion, the results of the constant and changing concentrati on experiments clearly show that protein binding reduces the in vitro antimicrobial activity. The study results further demonstrate that binding to commercially available protein supplements can substantially differ from that of serum or plasma and greatly depends on the supplement used. Correcting total concentrations for reported litera ture binding values is, consequently, unreliab le. Instead, free, active antibiotic concentrations should be experi mentally measured in the actual test system. In vitro microdialysis is a convenient sa mpling tool for this purpose. In addition, an international
35 standardization of the respective test systems might help prevent further misinterpretation of the impact of protein binding on the antimicrobial activity.
36 Table 2-1. Determined MIC values (presented as modes) and simultaneously fitted model parameters (SD) of ceftriaxone and ertapenem against E. coli and S. pneum. in the presence and absence of bovine serum albumin (BSA) and human serum albumin (HSA), respectively. Parameter (unit) Ceftriaxone vs. E. coli ATCC 25922 vs. S. pneum. ATCC 6303 w/out albumin w/ BSAa w/ HSAa w/out albumin w/ BSAa w/ HSAa MIC ( g/mL) 0.064 0.0642.00.010.02 0.16 ks (h-1) 2.40 (.34) 1.56 (.08) kmax (h-1) 6.34 (.13) 3.59 (.11) EC50 (SD) ( g/mL) 0.027 h (.0007) 0.057 h (.007) 1.096c, b (.00272) 0.004 h (.0004) 0.015 h (.0099) 0.084c, b (.0147) h 1.61 (.10) 2.18 (.14) MSC 4.24 4.04 a At a protein concentration of 40g/L. Significant differences (P 0.05) from the value for cthe control (protein-free), bthe sample with BSA and hthe sample with HSA.
37 Table 2-1. Continued Parameter (unit) Ertapenem vs. E. coli ATCC 25922 vs. S. pneum. ATCC 6303 w/out albumin w/ BSAa w/ HSAa w/out albumin w/ BSAa w/ HSAa MIC ( g/mL) 0.015 0.0150.030.0250.025 0.025 ks (h-1) 2.76 (.89) 2.28 (.13) kmax (h-1) 6.02 (.00) 3.28 (.11) EC50 (SD) ( g/mL) 0.010 (.0035) 0.020 (.0029) 0.014 (.0046) 0.010 (.0019) 0.010 (.0015) 0.019 (.0101) h 2.54 (.14) 2.65 (.25) MSC 4.71 2.67 a At a protein concentration of 40g/L. Significant differences (P 0.05) from the value for cthe control (protein-free), bthe sample with BSA and hthe sample with HSA.
38 Figure 2-1. In vitro protein binding of ceftr iaxone and ertapenem. In vitro mean protein binding (%) of ceftriaxone (white) and ertapenem (gray) in Todd-Hewitt broth (THB), THB with bovine serum albumina (BSA), THB with human serum albumina (HSA), pooled adult bovine serum (ABS) and pooled human plasma (HP), respectively aAt a protein concentration of 40g/L
39 Figure 2-2. Effect of bovine (BSA) and human serum albumin (HSA) on bacterial growth and antibiotic-induced kill. The maximum k ill-rate was determined at 8 times MIC (8xMIC, with respect to albumin-free medi um) for ceftriaxone and ertapenem against E. coli and S. pneum., respectively. (Symbols: without albumin ( ), with BSA ( ), with HSA ( ))
40 Figure 2-3. Simultaneous curve fits for ceftriaxone against E. coli and S. pneum. in the presence and absence of bovine serum albumin (BSA) and human serum albumin (HSA). At concentrations of 0.25xMIC ( ), 0.5xMIC ( ), 1xMIC ( ), 2xMIC ( ), 4xMIC ( ), 8xMIC ( ), 16xMIC ( ) plus growth controls ( ). Symbols represent the experimental data; solid lines represent the simultaneous curve fits based on the PK/PD model.
41 Figure 2-4. Simultaneous curve fits for ertapenem against E. coli and S. pneum. in the presence and absence of bovine serum albumin (BSA) and human serum albumin (HSA). At concentrations of 0.25xMIC ( ), 0.5xMIC ( ), 1xMIC ( ), 2xMIC ( ), 4xMIC ( ), 8xMIC ( ), 16xMIC ( ) plus a growth control ( ).Symbols represent the experimental data; solid lines represent the simultaneous curve fits based on the PK/PD model.
42 CHAPTER 3 CLINICAL MICRODIALYSIS IN SKIN AND SOFT TISSUES: AN UPDATE2 Introduction Historically, pharmacokinetic/pharmacodynami c (PK/PD) approaches link plasma drug concentrations to observed effects. However, mo st PD effects are mediated by interaction with enzyme, transporter or receptor systems that are located in the tissues (28, 44). Consequently, linkage of PD effects to tissue drug concentratio ns at the site of action is a more accurate approach to characterize exposureeffect relationships (44). Clos er evaluation of concentrationeffect relationships have shown that only free, unbound drug at the target si te is responsible for PD efficacy (13, 59). Therefore, continuous sampling of free drug in the tissues is the most rational approach to estimate active drug profiles at the site of action. MD is currently the most appropriate sampling technique that can provide this information (44, 120). The MD principle was first employed in the early 1960s in order to sample free amino acids and other electrolytes in the extracellular fl uid of animal brains (16). Further technical advancement resulted in the develo pment of the dialytrode in 1972, the first simple version of todays MD probe (66). In 1974, Ungerstedt and Pycock discussed the use of hollow fibers as a superior in vivo sampling technique (260). Steady im provement of both MD catheters and methodology allowed not only measurement of neurot ransmitters and metabolites in animals, but also application in humans (107). In early clinical tr ials, glucose levels were determined in subcutaneous adipose tissue (18, 63, 117). The approval of MD catheters for use in humans by the US Food and Drug Administration (FDA) and the European Union Conformite Europeene (CE) has opened the door to furt her studies in virtually every human tissue, including muscle, skin, lung, myocardium, brain, and even tumo rs (44). Consequently, in recent years, 2 Copyright American College of Clini cal Pharmacology, [J Clin Pharmacol 48: 351-64, 2008]
43 considerable experience has been gained in clinical studies of both healthy volunteers and patients, resulting in more than 2,000 publications as of today. Rationale for this Review In 2004, the Food and Drug Administration (F DA) issued its Criti cal Path Document, Innovation and Stagnation? Challenge and Opport unity on the Critical Path to New Medical Products (84). In this report, the FDA critically evaluates reasons for the recent decline of drug launches onto the market despite the increased in vestment of time and resources. One of the key issues identified in this document is the lack of sufficient safety and efficacy measures for new drugs and drug formulations (84). In the FDAs view, a new product development toolkit containing powerful new scientific and techni cal methods (i.e. animal or computer-based predictive models, biomarker for safety and effectiveness, and new clinical evaluation techniques) is urgently needed (84). Recently, the FDA published a list of Critica l Path Opportunities. (81) MD is one of these new clinically applicable evaluation technique s that is specifically mentioned in this list. Its capability of determining free, active local drug concentrations qualifies it for site-specific safety and efficacy assessment. This has become of particular interest to industry and regulatory authorities for the evaluation of bioavailability (BA) and bioequivalence (BE) of topically applied drug formulations, especially of derm atological products (81, 82, 85). For systemic drugs, usually the serum concentrations are used in BA and BE studies. However, based on the respective definitions of BA and BE, measurements of exposure at the site of action would be more meaningful (83). The goal of our review is to give an overview of newer clinical BA and BE studies in skin and soft tissues using microdialysis.
44 Methods The MD catheter (probe) consists of a small semi-permeable hollow fiber membrane that is connected to inlet and ou tlet tubing (63). It is constantly perfused with a physiological solution (perfusate) at flow rate s of approximately 0.1-5 L/min (44). After insertion into a selected tissue or (body) fluid, solutes can cross the me mbrane by passive diffusion depending on their concentration gradient (28, 44, 63). Hence, the pr obe can be used as a sampling tool as well as a delivery tool (44). The solution l eaving the probe (dialysate) is co llected at certain time intervals for analysis. Calibration Methods Since the MD probe is conti nuously perfused with fresh pe rfusate, a total equilibrium across the membrane cannot be established. Rath er, a steady-state rate of exchange across the MD membrane is rapidly reached. This steady-st ate exchange rate is described by the extraction efficiency (EE). The EE is the ratio between the loss/gain of analyte during its passage through the probe (Cperfusate-Cdialysate) and the difference in concentr ation between perfusate and the sample of interest such as tissue fluid, in vitro analyte, etc. (Cperfusate-Csample), as shown in equation 3-1. sample perfusate dialysate perfusateCC CC EE (3-1) At steady-state, EE has the same value for all Cperfusate, no matter if the analyte is being enriched or depleted in the perfusate (44). Fo r this reason, MD probes can be calibrated with either drug-containing perfusate or sample solutions. While various calibration techniques are available (i.e. low-flow-rate method, zero-netflux method (161), extend ed zero-net-flux method (207), etc.), retrodialysis by drug (34) is th e most commonly employed method in humans.
45 During retrodialysis, the probe is perfused with drug-containing perfusate prior to or after drug administration, without the drug in the tissue. Since abse nce of drug in the tissue is required for retrodialysis, this calibration technique cannot be applie d to endogenous compounds (44). The proper selection of an appropriate calibra tion method is critically important for the success of a MD experi ment, so supportive in vitro experiments prior to use in animals or humans are recommended (44). The recovery determined in vitro might differ from the recovery in humans, therefore, its actual value n eeds to be determined in every single in vivo experiment (241). Strengths of the Microdialysis Technique Depending on the molecular cut-off of the MD probe membrane, larger molecules (such as proteins) are prevented from diffusing into the di alysate (241). This allows the analysis of protein-free, active drug concentr ations to be performed frequently without further sample preparation (28). While MD selectively determines free, unbound con centrations in the interstitial fluid of a particular living tissue, other sampling techniques have limited capabilities to distinguish between different sites within the tissue or between free and bound drug. For example, tape stripping is a commonly used met hod that is well esta blished for evaluating the penetration of topically applied compounds into the upper part of the skin (150) However, this method is limited to the stratum corneum and can, therefore, not be used for the assessment of free, active drug concentrations in deeper tissues such as the dermis. In fact, continuous tape stripping does disturb the barrier function of the skin and can re sult in artificially increased drug levels in the skin (15). Other tissue sampling approaches al so have major limitations. While, for example, broncho-alveolar lavage (BAL ) pools data from large segments throughout the lung, concentrations obtained from homogenized ti ssue, positron emission tomography (PET), and
46 scintigraphy will include drug that is bound to inters titial and intracellular pr oteins or to intraand intercellular membrane structures (28, 44, 241). In comparison, the skin blister technique is capab le of target-site specific sampling. Yet, it reflects concentrations in experimentally induced secretory fluids The concentrations in these fluids might vary with blister size and surface to volume ratio due to protein and chemokine content and might not be comparable to the inte rstitial space fluid (28, 48, 61, 189, 222, 241). It was shown, for example, that ciprofloxacin a nd moxifloxacin accumulated preferably in blister fluid whereas an almost complete equilibration of the free unbound antibiotic plasma fraction with the interstitial space fluid was observed using MD (28, 29, 193). Another limitation of most of these techniques is that they are usually not capable of continuous measurement of the c oncentration-time profile. Hen ce, investigators using, for example, epithelial lining fluid (ELF) and tissue bi opsies are forced to pool data from different subjects in order to receive concentration-tim e profiles (1, 278). In contrast, continuous sampling via MD allows the generation of PK profiles from individual subjects. Limitations of the Microdialysis Technique Usually, MD probe insertion is associated w ith minimal tissue damage. However, some tissue sites such as brain (78, 79), lung (102, 254), bone (253), heart (9), liver (204) or the peritoneal cavity (116) are not readily accessible to the MD procedure. The MD probe must then be surgically implanted into these tissues. Another limitation of the MD technique is it s dependency on flow rate and sensitivity of the analytical assay. As the analyte concentra tion decreases with increasing flow rate, assays with small sample volumes are restricted to hi ghly sensitive analysis techniques (i.e. HPLC, LC/MS/MS, capillary electrophoresis) (44). However, if very low flow rates are used, the time
47 resolution might be compromised (44). For this reason, flow rate and analytical procedures require extensive fine-tuning. MD has emerged as the method of choice to monitor drug concentrations in the extracellular space. If the site of action is located intracellularly, MD is not able to measure that concentration directly. Even in these cases, th e respective extracellula r concentration resides closer to the site of inte rest than the respective plasma or blood concentrations. Bioavailability at the Site of Action In the FDA Guidance for Industry, BA is defi ned as the rate and extent to which the active ingredient or active moiety is absorbed from a drug product and becomes available at the site of action. For drug products that are not intended to be ab sorbed into the bloodstream, BA may be assessed by measurements that reflect the rate and extent to which the active ingredient or active moiety becomes available at the site of action. (83) Lo cal site drug levels can become important in, for example, anti-cancer therapy, because tumors might show altered physiology and/or limited drug access compared to normal ti ssue (114, 115, 213). Knowledge of the active drug concentrations inside the tumor is therefor e critically important for the selection of an appropriate drug and/ or dosing regimen. Even though metastatic malignant melanom as (MMMs) respond poorly to drugs due to resistance at a molecular level and impaired tr anscapillary drug transf er, dacarbazine is an effective treatment of MMMs (121) Transcapillary tran sfer rates of dacarbazine and its active metabolite 5-aminoimidazole-4-carboxamide (AIC) into the tumor were determined after intravenous administration of dacarbazine at doses of 200mg/m2 -1000mg/m2 body surface area in 7 MMM patients using MD. Dialysates (tum or and healthy adipose tissue) and plasma (ultrafiltered) samples were collected ove r 240min and analyzed using HPLC for free dacarbazine and AIC concentrations. Results indicated that for all doses, AUCs for dacarbazine
48 and AIC were not significantly different betw een plasma (free concentration) and tumor interstitium. It was concluded that dacarbazine and its active metabolite AIC showed significant tumor penetration characteristics after i.v. admini stration. The lack of response to antineoplastic therapy with dacarbazine might be explained by resistance at molecular level rather than by inability of dacarbazine and AIC to pene trate into the interstitium of MMM (121). Once the drug reaches the systemic circulat ion, there are no more differences in distribution, metabolism, and elimination be tween the intravenous and the non-intravenous administration route. Depending on the value of th e oral bioavailability (F), oral doses will have to be increased in comparison to the respective intravenous dose in or der to achieve similar therapeutic drug levels. As a case in point, MD was employed to compare free, active ciprofloxacin concentrations in the interstitial space fluid (ISF) of skeletal muscle and subcutaneous ad ipose tissue, after i.v. or oral ciprofloxacin ad ministration, respectively (29). Each of eight healthy volunteers were studied twice and randomly assigned to initial ciprofloxacin treatm ent with either 500mg orally or 400mg intravenously, with a washout period of at least 7 days. Free ciprofloxacin concentrations were determined in the ISF of sk eletal muscle and subcutaneous adipose tissue, saliva, cantharis-induc ed skin blister, as well as capillar y plasma, and compared to total venous plasma concentrations. Samples were anal yzed using HPLC, and respective AUCs were calculated. In order to predict the antimicr obial activity of ciprof loxacin, PK profiles, determined in the ISF after oral and in travenous dosing, were simulated in an in vitro PD model against Enterobacter, K. pneumoniae, and S. aureus. Results showed that after oral and intravenous administration, mean fAUCs ( SD) of both muscle a nd subcutaneous adipose tissue were statistically significantly lower than the corresponding AUC for plasma (29). Whereas
49 fAUCmuscle i.v. (7.43 1.40 mgh/L) was significantly higher than fAUCmuscle oral (4.49 1.41mgh/L), no significant differences could be detected between fAUCsubcutis i.v. (4.13 1.63mgh/L) and fAUCsubcutis oral (3.85 2.26mgh/L) (29). However, a closer look at the continuously increasing C muscle oral/C plasma ratio indicates that stea dy-state conditions have not yet been reached. While a Cskin blister/Cplasma ratio > 4 is an indicator th at ciprofloxacin preferably accumulates in inflamed lesions, saliva and capilla ry blood concentrations were similar to total plasma (29). In addition, results from the in vitro PD model showed that a comparable outcome was achieved against selected test strains with ciprofloxacin give n either as 400mg i.v. or 500mg orally (29). The above information leads to the conclusion that single i. v. infusion of 400mg and single oral administration of 500mg of ciprofloxacin result in different skeletal muscle and subcutaneous adipose tissue concentrations. Yet, these differences in PK might not be pronounced enough to result in clinically si gnificantly altered PD outcome (29). In some cases, on the other hand, the absorp tion of active drug mo lecules into the blood stream is undesirable, especially when drug-sp ecific systemic advers e events are induced. Topical administration of these compounds might, theref ore, be considered as an alternative. For example, nonsteroidal anti-inflammatory drug s (NSAIDs) such as diclofenac are widely prescribed for the treatment of rheumatic diseases (105). Although they are among the most commonly prescribed drugs worldwide, they are also responsible for approximately one-quarter of all adverse drug reactions such as increased ri sk of severe gastrointestinal (GI) complications (105). Topical diclofenac administration might th erefore be superior to oral or intravenous application for the treatment of inflammatory diseases. MD was used in the following study to determ ine the relative BA of diclofenac in plasma, subcutaneous adipose and skeletal muscle tissue af ter application of a novel diclofenac spray gel
50 formulation (4%), or oral dosing w ith enteric coated diclofenac tabl ets, respectively (25). In one study, 12 healthy, male volunteers received two dosing regimens with a 14-day washout period in between. During the first regimen, a diclof enac spray gel formulation (48mg) was applied topically TID for three days (10 doses total). In the 2nd regimen, enteric-coated diclofenac tablets (50mg) were administered orally TID for three days (10 doses total). In both cases, after administration of the 10th dose, blood and MD (from subcutaneous adipose and skeletal muscle tissue) were collected in 1-hour intervals for 10 hours post dose and at 48h. Diclofenac concentrations in dialysate and plasma samp les were determined using LC-MS-MS and the respective AUCs calculated. While the relative BA of diclofenac in skeletal muscle (209%) and subcutaneous adipose tissue (324%) was higher after topical compared to oral administration, the relative BA in plasma was 50-fold lower (Table 3-1) (25). In addition, results showed that maximum plasma concentrations (Cmax,plasma) after topical administration were approximately 250-fold lower than the Cmax,plasma values after oral administrati on. These data lead to the conclusion that the spray gel form ulation is a good alternative to oral diclofenac formulations for the treatment of inflammatory soft tissue conditions due to the favorable penetration characteristics and low systemic availability (25). Although the results of the previous study i ndicate topical diclofenac formulations are appropriate for treatment of rheumatic diseas es, there is still uncertainty whether these concentrations are sufficient in the target tissues (215, 217, 235) The aim of another MD study was to investigate transdermal penetration of di clofenac after local topical administration into superficial and deep tissue laye rs (191). Two MD catheters one into a superficial (3.9 0.3mm) and one into deep (9.3 0.5mm) were inserted into the tissue of 20 healthy, male volunteers. The correct position (d istance between skin surface a nd MD tip) was determined by
51 high frequency ultrasound scanning. Diclofenac gel (approximately 300mg/100cm2) was applied onto skin above the inserted MD membrane. Di alysate was collected every 30 min for up to 4 h and respective free AUCs were calculated. Results showed that diclofenac concentrations could be determined in both sampling sites in just 7 volunteers and could not be correlated to the insertion depth. Linkage to the IC50 (0.5 g/mL) of cyclooxygenase 2 (COX-2) showed that effective levels in underlying tissue layers were reached in only 8 out of 20 subjects (179, 191). In contrast to the previous study, it was therefore concluded that due to insufficient deep tissue penetration, a generalized use of transdermal diclofenac formulati ons, at least single doses, is often not justified and may be greatly depe ndent on individual sk in properties (191). The effect of NSAIDs is triggered by inhibi tion of key enzymes (cyclooxygenase 1 and 2) of the prostaglandin (PG) synt hesis. PGs are mediators a nd are involved in a number of physiological and pathophysiological effects including the evolvement of pain (31). However, it is still debated whether antihyperalgesic effects of NSAIDs include both peripheral (inflammation site) and central sites of action (31). MD was used to determine PG levels in the skin after topical and oral diclofenac administration. All of the 10 healthy volunteers were treated in three consecutive treatment periods. Each treatment period was randomly assigned and could contain the following combinations: a) oral formulation (93mg) plus to pical formulation (placebo) b) oral formulation (placebo) plus topical formulation (65mg) or c) oral formulation (placebo) plus topical formulation (placebo), respectively. While MD samples were taken every 0.5h after administration for up to 6 h, blood samples were drawn for 24h and analyzed using LC-MS-MS. In addition, antihyperalgesic ac tion of diclofenac was assessed using an inflammatory model of cutaneous hyperalgesia (freeze lesion). The response was quantified by estimating mechanical
52 pain threshold (MPT) before and after dosing at 0.5h intervals for up to 6h. Study results showed that both topical and oral diclofenac formulations are si gnificantly more effective than placebo. While higher tissue levels were measur ed for topical treatment during the first hour after the application compared to oral administration (46.1ng/mL vs. 11.4ng/mL), oral administration resulted in higher tissue levels at time points 2 and 2. 5h (Figure 3-1) (31). However, after topical administ ration, diclofenac could not be detected in plasma. Even though the area under the curve in tissue (fAUCtissue) after oral dosing was lower than after topical application (32.2ngh/mL vs. 40.7ngh/mL), the overall pain relief was 1.7 fold higher after oral administration during th e first three hours. Accordingly, the authors suggested that an additional centrally mediated antihyperalgesic eff ect is involved in the analgesic effect of systemically administered diclofenac (31). Insufficient BA at the tissue site, in additi on to the occurrence of adverse events, might limit the use of oral or IV formulations (250) Instead, topical administration can overcome these limitations and is employed to deliver drugs at, or close to, the point of application (221). Topical drug formulations can be applied to th e eye, nose and throat, ear, vagina, lung, etc.; however, the vast majority of topical medications is applied to the skin (221). Dermal drug formulations are thus commonly used for the trea tment of skin diseases such as urticaria, psoriasis or skin cancer. In the area of skin cancer treatment, phot odynamic therapy (PDT) has shown promising results for the treatment of basal cell carcinoma (BCC), the most common type of non-melanotic skin cancer (266). During PDT therapy, an intravenously or t opically administered photosentisizer such as Protopor phyrin IX accumulates in the cancer tissue (266). After exposure to light of a specific wa velength, this photosensitizer releases cytotoxic singlet oxygen.
53 MD was used to assess the BA of Delta-Aminolevulinic Acid ( -ALA), a prodrug, that is specifically metabolized by tumor cells to Protop orphyrin IX, in BCC (n = 14) and normal skin (n = 4) after dermal application (266). In a ddition, the skin blood flow was mapped and skin amino acid content determined using laser Doppl er perfusion imaging and MD, respectively. Results indicated that inte rstitial ALA concentration in BCC increased from 0 to 3.1 1.7mmol/L (mean SEM) within 15 min of application, whereas no ALA could be detected in healthy skin. In contrast, amino acid levels were found to be similar in both healthy and BCC tissue and blood flow was 2.5-fold increased in BCC compared to normal skin during treatment (266). It was concluded that MD is an appropriate tech nique for the determination of ALA PK in the skin. However, the rapid pe netration of ALA into tumor tissue and increased blood flow in BCC might lead to faster elimination from th e tumor and warrants further investigation. Factors Affecting Bioavailability Despite great interest in the skin as an application route for therapeutic agents, the availability of clinical BA studies evaluati ng the mechanisms of tr ansdermal absorption and factors affecting the disposition of topically applied drugs is lim ited (15, 21, 55). In order to address this lack of information, an increasi ng number of human studies have been performed evaluating factors (e.g. skin barrie r function, blood flow, degree of ionization, etc.) that can alter the BA after dermal drug application. The ability of MD to continuously monitor the change of free, unbound drug in the ISF of diffe rent layers of the skin or subcutaneous adipose tissue has made it a valuable tool for inves tigating these factors. MD has, for example, been used to study the effect of skin barrier pert urbation by repeated tape stripp ing, treatment with 1% sodium lauryl sulfate (SLS) or 2% SLS, and treatment w ith acetone on the penetration of salicylic acid into the skin (15). Findings show that ther e was an approximate 150-fold difference between
54 unmodified and tape-stripped/SLS (2%) treated skin, indicating a ma ssive disturbance of the skin barrier function (Figure 3-2). In comparison, pe netration increased appr oximately 46-fold after SLS (1%) and 2.2-fold after acetone treatment, respectively. The BA of topically applied drugs in skin and soft tissue is not only dependent on the integrity of the skin barrier but also highly corre lated to the local blood flow (17, 19, 21, 44, 119, 219, 234). An increase, as well as a decrease, in local blood flow can be due to physiological or drug-induced changes. Whereas an increase in blood flow enhances the penetration of a compound into the respective tissue, a decrease slows down its tissue uptake (19, 21, 30, 44, 119). This situation was demonstrated in a MD st udy evaluating the BA of penciclovir (PCV) in the skin after a single oral dos e (250mg) of its prodrug famciclovi r (19). Three MD probes were implanted into the left forearm of seven healthy volunteers. While all th ree probes were perfused with Ringer solution, vasoconstriction was a dditionally induced in th e tissues around probes #2 and #3 by either supplementation with ad renaline (0.2mg/mL) or cooling (20oC), respectively (19, 108). Results (Figure 3-3) indicate that penetration of PSV into untreated skin was statistically significantly higher than into skin with decreased microcirculation due to either adrenaline or cooling, respectiv ely (19). In comparison, loca l warming of the dermis (40oC) was shown to enhance the microcirculation in soft tissues and was correlated with an increased penetration of ciprofloxaci n into these tissues (119). However, physiological or pathophysiological conditions determine not only the BA of a drug in the soft tissue, but the physicochemical pr operties of the drug itself. It was shown in both in vitro and in vivo experiments that unioni zed drug penetrates more efficiently through the stratum corneum than ionized drug (145). Whereas chemical modifications of the drug molecule
55 (e.g. ester, ion pairs, etc.) are frequently em ployed to increase the lipophilicity of the parent compound, the molecules charge can also be acti vely used (e.g. in iont ophoresis) for active drug delivery (55, 240). Yet, these chemical modificati ons can result in different penetration behavior and consequently altered relative BA (55). Alte red BA can become an issue when two different drug formulations of the same parent compound ar e used and can be addr essed in BE studies. Bioequivalence In the case of systemically active drugs, alt hough both BA and BE evaluate the release of a drug substance from a drug product and the subseque nt absorption into the systemic circulation, BE is a more formal comparative test between tw o drug products using specifi ed criteria. In the FDA Guidance for Industry, BE is defined as the absence of a significant difference in the rate and extent to which the active ingredient or active moiety in pharmaceutical equivalents or pharmaceutical alternatives becomes available at th e site of drug action when administered at the same molar dose under similar conditions in an appropriately designed study. (83) However, the determination of BE for locally acting and targeted delivery products has confronted both industry and regulatory authori ties with problems during the a pproval process since plasma concentrations are usually inappropriate surrogates of pharmacological activity (81). A recent study compared the applicability of MD for BE determination of two topical lidocaine formulations (cream 5%, ointment 5%) to the tape stripping method (14). Multiple MD and tape stripping samples were taken at two different application sites. Results of both methods showed that these two formulations we re not bioequivalent, and consequently, not interchangeable. In addition, further statistical an alysis of the applied st udy design indicated that BE studies (90% CI and 80-125% BE limits) usi ng two formulations in each subject need a minimum of 27 subjects (when 2 probes are used per application site), or 18 subjects (when 3 probes are used per application site ), respectively (14). In comparison, it has been estimated that
56 approximately 40 to 50 subjects are required for BE studies using the tape stripping method and up to 300 subjects, using the current clinical BE study design (14, 44). PK/PD Indices Besides assessment of BA and/ or BE, the actual informati on on free drug concentrations obtained from the MD experiments can be furthe r used to predict treatment outcome. This approach is frequently employe d during drug development of an ti-infective agents. In the following clinical MD studies, measured free, active antibiotic concentrations were linked to the respective PD outcome parameters of the most pr evalent skin and skin structure pathogens in order to predict their clinical efficacy. Infections of skin and soft tissue can be cau sed by a variety of gram-positive and gramnegative pathogens and are routinely treated with antibiotics. Wh ereas penicillins and cephalosporins are drugs of first choice, agen ts of different classe s (e.g. oxazolidinones, glycopeptides, macrolides, tetracyclines, etc.) have to be used in case of adverse events or emergence of -lactam resistance. In order to increase the chances of clinical success and to decrease the likelihood of toxic si de effects as well as resistance development, selection of an appropriate antibiotic dosing regimen is extremel y important (229). The most rational approach is to link active drug concentrations to th e respective PD outcome. However, outcome predictions based on total plasma concentrations might be misleading since most infections are not located in the blood stream, but rather in the interstitial space fluid (ISF ) of tissues (193). In fact, it is free, unbound drug in the ISF that is re sponsible for antimicrobial efficacy (222). Once free antibiotic concentrations ha ve been determined at the in fection site, outcome should be predictable from the respective susceptibility breakpoints of the infection-causing pathogens (26).
57 To date, the minimum inhibitory concentration (MIC) has served as a well-established and routinely determined susceptibility breakpoint parameter for antibiotics According to the Clinical and Laboratory Standards Institute (C LSI, formerly National Committee for Clinical Laboratory Standards NCCLS), the MIC is defined as the lowest concentration of drug that completely inhibits visible growth of the organi sm as detected by the unaided eye after an 18-24 hour incubation period with a standa rd inoculum of approximately 5x1056CFU/mL (188, 229). Combinations of this PD marker with free (), unbound PK parameters to MIC-based PK/PD indices such as T>MIC, AUC/MIC and Cmax/MIC have led to a much better understanding of antib iotic dosing (229). The first PK/PD index was developed for penicillins. It correlates in vivo efficacy with the amount of time free drug levels stay above the MIC of the target organism (T>MIC) (229). Although a further index differentiation with in the drug class was suggested, a common threshold of T>MIC 40% seems to be sufficient for the clinical efficacy of -lactam antibiotics (229). While Cmax/MIC index values of 10-12 seem to be a good predictor for aminoglycosides, the magnitude of the fluoroquinol one index is still cont roversial (89, 125, 180, 214). Nonetheless, target AUC24/MIC values of 100-125 (Gram-negatives), 25-35 (Grampositives) and Cmax/MIC index values of 10 have been id entified for fluoroquinolones (76, 89, 205). In comparison, AUC24/MIC values 50-100 or T>MIC 85% were good outcome predictors for oxazolidinones ( 168). Once these MIC-based PK/PD indices are identified, they can support the identification of optimised dosing regimens and the prediction of treatment outcome (229). Knowledge of the free antibiotic concentration-tim e course in the ISF is necessary in order to establish the respective T>MIC, Cmax/MIC and AUC24/MIC index values. Whereas various
58 techniques are available for determination of free, unbound concentrations, they are not all capable of characterizing dynamic changes in free ISF concentrations. Only MD combines these two properties. Consequently, it is therefore a very valuable sampling tool and has become an inherent part of evaluation and establishment of PK/PD indices. T>MIC Although clinical studies have demonstrated the effectiveness of ertapenem in SSSI treatment, few studies on the in vivo penetration of ertapenem into ISF of soft tissues, such as skeletal muscle and subcutaneous adipose tissue, and resulting free, active concentrations, have been available (34, 99, 151). In a single center, prospec tive, open label study, free, unbound ertapenem concentrations in the ISF of skeletal muscle and subcutaneous adip ose tissue were measured using MD (34). After determination of the individual pr obe recoveries, six healthy volunteers received 1g ertapenem as a single 30 min short-term i.v. infusion. Plasma and MD samples were collected over 12 h postdosing and analyzed using LC-MS-MS. Results indicate that free, unbound ertapenem profiles in the ISF of both skeletal muscle and subcut aneous adipose tissue ar e lower than corresponding total plasma concentrations as shown in Figure 3-4. While free ISF concentrations of the skeletal muscle correlated well with free, unbound concentr ations in plasma (4-16% of total plasma concentration), they were comparably higher th an free ISF concentrations in subcutaneous adipose tissue. This phenomenon was observed in other studies as well, and might be explained by differences in blood flow in these two tissues (27). Free ertapenem concentrations of 1.13 0.68mg/L in the muscle, observed 12h after single dose i.v infusion of 1g ertapenem, exceeded the MIC90s of methicillin susceptible S. aureus (0.25mg/L), Streptococcus spp (0.5mg/L), extended-spectrum b-lactamase (ESBL)-producing Enterobacteriaceae (0.03-0.06mg/L), Bacteroides fragilis and other anaerobic bacteria ( 1.0mg/L) for at least 50% of the entire
59 dosing interval. In comparison, free levels of 0.31 0.16mg/L in the subcutaneous adipose tissue (at 12h) exceeded the MIC90 of the same SSSI pathogens for at least 30% of the dosing interval. The authors concluded that free, act ive ertapenem concentrations reached sufficient levels in non-infected interstiti al fluid of muscle and subcutaneous adipose tissue and that previous clinical findings were supported by this data (34). AUC24/MIC Linezolid, the first oxazolidinon e, is approved by the FDA fo r the treatment of nosocomial pneumonia and complicated skin and skin structure infections It shows good antimicrobial activity against various resistant gram-positive bacteria, including methicillinand glycopeptideresistant S. aureus (282). Despite the fact that 1) only free, unbound data is considered for antimicrobial efficacy and 2) most relevant pa thogens are located in the ISF, most of the available linezolid PK data is ba sed on total plasma concentrations (33, 64, 243). Therefore, a clinical MD study was performed th at evaluated the penetration of linezolid into soft tissues of 10 healthy volunteers after single and multiple do se administration (64). On day one of this study, MD catheters were placed into the subcutan eous adipose tissue and the skeletal muscle of each volunteer. After calibrati on and baseline determination, 600mg linezolid were infused intravenously over 30min. MD and blood samples were taken for up to eight hours. After withdrawal of the MD probes, volunteers were started on oral linezolid (600mg) BID for five consecutive days (64). The second set of MD experiments was started simultaneously with the last oral dose. This time, probes were calib rated after the 8-hour sampling period. The AUC0-8 was calculated by using the trapezoidal rule, the AUC24 by extrapolation to 24 hours. In addition, AUC24/MIC ratios were calculated for pathoge ns with MICs of 2mg/L and 4mg/L, respectively. Results show that after si ngle i.v. administration of linezolid, AUC0-8 of both skeletal muscle (65.3 18.2mgh/L) and subcut aneous adipose (75.8 24.2mgh/L) tissue were
60 statistically significantly higher than the AUC0-8 of plasma (53.0 11.6mgh/L). However, at steady-state no significant differences between concentrations in the ISF of skeletal muscle and subcutaneous adipose tissue could be detected (6 4). The findings furthe r indicated that steady state concentrations in both muscle (AUC24 muscle/MIC 58.9 33.0mgh/L) and adipose subcutaneous tissue (AUC24 tissue/MIC 46.6 15.9mgh/L) were sufficient to treat infections that are caused by pathogens with MICs of up to 4 mg/L (64). Cmax/MIC According to the CDC (Center of Disease Contro l), surgical site infections (SSI) include skin and subcutaneous tissue infections and are the second most common cause of serious nosocomial infections (23, 195). These SSI can be caused by various pathogens (e.g. S. aureus, P. aeruginosa, Klebsiella ssp.) and are frequently treated with -lactam antibiotics (195, 263). In patients with confirmed alle rgy or adverse reaction to -lactam antibiotics, gentamicin can be used in combination with either clindamycin or metronidazole (24). Gentamicin is a broadspectrum antibiotic that shows good antimic robial activity against gram-positive S. aureus and gram-negative bacteria including Enterobacteriaceae and Pseudomonaceae (162). Its short halflife of 2-3 hours in patients with normal renal function requires a well-designed dosing regimen in order to prevent concentrations from dropping rapidly to subtherape utic levels (24). Since it is oftentimes unclear whether surgical procedures change free, active drug levels in the ISF, the most rational approach is to measure respective fr ee concentrations directly at the surgical site using MD. After insertion of the MD catheter into the subcutaneous fat layer of the abdominal wall (10cm lateral to the umbilicus), seven healthy vo lunteers, with normal renal function, received 240mg gentamicin as an IV bolus (162). For the first hour, plasma and MD samples were taken every 20min, followed by 60min sampling intervals for up to six hours. Samples were analyzed
61 using a spectrophotometric immunoassay. PK parameters were determined by compartmental analysis. Free peak concentrations in tissue (Cmax tissue) of 6.7 2.0mg/L, which is equivalent to 39.1% of the total peak serum concentration (Cmax), were reached within 10 to 30min of administration (162). Since the Cmax/MIC ratio was identified as the best predictor of aminoglycoside efficacy, Cmax/MIC values were calculated for common SSI pathogens such as P. aeruginosa (7.4:1, MIC: 0.9mg/L), S. aureus (33.5:1, MIC: 0.2mg/L) and Klebsiella ssp. (4.8:1, MIC: 1.4mg/L), respectively (162). Th e authors concluded that 1) MD could be successfully used to measure gentamicin con centrations in subcutaneous tissue and 2) determined gentamicin concentrations were suffi cient to treat infections with the most common SSI pathogens (162). In some cases, the antibiotic dosing regimen does not result in sufficient concentrations in the ISF. If they fail to outreach the MIC thre sholds of respective SSSI pathogens, they cannot be used for treatment. Telithromycin, for example, is typically used as a reserve antibiotic for the treatment of respiratory tract infections since it shows hi gh concentrations in inflammatory fluids, bronchopulmonary tissues, tonsillar tissue, and saliva (74, 97, 128) However, its good activity against some Streptococci ssp. has lead to speculation on its applicability in the treatment of SSSI (94, 97). Hence, a MD study was designed to evaluate the PK profile of telithromycin in the ISF of soft tissues after single dose administ ration. In this study, commercially available 800mg telithromycin tablets were given orally to 10 healthy male volunteers (94). Blood and dialysate concentrations were determined. Ar eas under the concentratio n time curve from zero to eight hours (AUC0-8) were calculated for free concentrat ions in the ISF of muscle and subcutaneous adipose tissue as well as free and total plasma concentrations, respectively.
62 Results showed that there were no statistically significant differences between areas under the mean free concentration-time curve from zero to eight hours (AUC0-8) in muscle (0.6 0.3mgh/L), subcutaneous adipose tissue (0.9 0.6mgh/L) and plasma (0.5 0.2mgh/L). However, AUC0-8s of muscle and subcutaneous adipose tis sue were significantly lower than the mean AUC0-8 of total plasma (4.1 1.5mgh/L) as shown in Figure 3-5. Since antimicrobial efficacy of ketolides correlates best with AUC0-24/MIC ratio, the AUC0-24 was additionally calculated. Using the MIC where the growth of 90% of the SSSIcausing pathogens is inhibited (MIC90), the AUC0-24/MIC90 ratios indicated that bacteria highly susceptible to telithromycin such as S. pyogenes might be eradicated from tissues and plasma (54, 276). Nevertheless, these AUC0-24/MIC90 ratios further indicated that telithro mycin shows insufficient concentrations against other prevalent SSSI pathogens such as S. aureus or bite pathogens such as Prevotella canis. The authors concluded, that telithromy cin shows limited activity against common SSSIcausing pathogens and should therefore not be used for their treatment (94). Conclusion Despite the fact that most pharmacological even ts take place in tissues, PK data is usually still based on blood or serum concentrations (44). However, free, unbound drug concentrations can differ significantly between blood and tissues. Measurement of free, active drug concentrations in the tissue of interest is consequently the most ra tional approach, but has frequently been restricted by insufficient sa mpling/analysis techniques. With MD, a wellaccepted, relatively inexpensive, semi-invasive samp ling technique that is able to sample free, active drug directly from the extracellular space fluid of tissues has become available. The resulting tissue concentration-time profiles reflec t both rate and extent of drug absorption from the respective application site and can, therefore, be used for BA and BE assessment. In fact,
63 MD has been recommended by the FDA as a sa mpling tool for BE evaluations of topical dermatological products (14, 44, 81, 134, 135, 173). To date, most of the current BA studies ar e performed in healthy volunteers. However, physiological processes can be altered in patient populations compared to healthy volunteers. Consequently, the assessment of BA in patients can be expected to become more important in the future. Initial progress has already been made in this area by determining free, active drug concentrations in diseased tissue (e.g. tumors, diabetic food, psoriasis, etc.) using MD. Once free, active drug concentrations have been determ ined, they then can be further correlated with appropriate biomarkers in orde r to predict clinical efficacy.
64 Table 3-1. Pharmacokinetic parameters for diclof enac in plasma and subcutaneous and skeletal muscle tissue. Parameters Topical Oral Median 95% Cl Median 95% Cl Plasma (n=12)a AUC AUC (nghmL-1) 32.8(22.7-52.9)1569.7 (1255.8-1849.8) Cmax (ng/mL) 4.9(3.4-7.7)1240.2 (787.0-1388.9) Subcutaneous tissue (n=12) b AUC AUC (nghmL-1) 21.5(19.4-50.5) 8.6 (7.0-10.6) Cmax (ng/mL) 13.1(9.3-33.6) 1.9 (1.6-2.5) Skeletal muscle (n=12) b AUC AUC (nghmL-1) 18.2(11.8-28.1) 8.8 (7.8-12.3) Cmax (ng/mL) 12.3(6.2-22.0) 2.6 (2.0-4.0) Data from 12 healthy males after multiple dose regimens of either topical (diclofenac spray gel 4%) or oral (50mg enteric-coated tablets) diclofenac. CI = confidence interval; AUC AUC = area under the plasma or tissue concentration vs time curve of diclofenac approximated to infinity (AUC ) or evaluated in the last dosage interval (0-8h; AUC); Cmax = maximum plasma or tissue concentration. aTotal drug. bFree drug
65 Figure 3-1. Mean concentration-tim e profile of diclofenac after oral and topical administration in subcutaneous tissue and plasma.
66 Figure 3-2. Concentration (mean SD)-time plots demonstrating differences in dermal salicylic acid (SA) penetration sampled by MD probes, inserted in the 4 barrier-pertubated skin areas (n=16).
67 Figure 3-3. Median concentration-time profiles for penciclovir (PCV) in skin for control, solution perfused with adrenaline, and cold skin fo llowing single oral ad ministration of 400mg famciclovir (prodrug) in healthy volunt eers (n=4).
68 Figure 3-4. Ertapenem concentration (mean SD)-t ime profiles in total plasma, skeletal muscle fluid, and interstitial adipose tissue (subc utis) following 1g infusion for 30min in healthy volunteers (n =6). Horizontal lines indicate MIC90 values for methicillin susceptible S. aureus ( ), Streptococcus spp ( ), extended-spectrum blactamase (ESBL)-producing Enterobacteria ceae (), Bacteroides fragilis and other anaerobic bacteria ().
69 Figure 3-5. Telithromycin concentration-time profiles in plasma (total and free), muscle, and subcutaneous adipose tissue after a singl e 800mg dose in healthy volunteers (n=10).
70 CHAPTER 4 PENETRATION OF ERTAPENEM INTO SKELETAL MUSCLE AND SUBCUTANEOUS ADIPOSE T ISSUE IN HEALTHY VOLUNTEERS MEASURED BY IN VIVO MICRODIALYSIS3 Introduction In clinical practice, empiric antibiotic therapy is initiated as the first measure dealing with bacterial infection. In case of therapeutic fa ilure, however, it may become necessary to measure the susceptibility of the causative agent and to adjust antibiotic dosage schedules with the aim to attain serum drug concentrations above the MI C for the respective pathogen throughout the dosing interval (53). Although this approach is suitable fo r conditions, where the central compartment is the main site of infection, e.g. septicemia, for localized organ or tissue infections drug concentrations in the interstitial space rather than in serum determine the clinical outcome of antimicrobial therapy (152, 188) This is particularly releva nt for skin and skin-structure infections (SSSI), which may occur due to surg ical wound contamination, after trauma or in diabetic patients and may result in severe necr otizing limband even life-threatening infections (32, 92, 96). Hence, to be clinically effective an antibiotic should reach pharmacologically active, i.e. unbound, soft tissue concentrations high enough to er adicate the causative pathogen (136, 177). A class of antibiotic s which was shown to qualify for the treatment of SSSI are carbapenems. Ertapenem is a long-acting parenteral 1-methyl-carbapenem, which was selected for clinical development partially ba sed on its pharmacokinetics (56, 156, 281). Owing to its long plasma half-life, which reflects a high plasma protein binding, ertapenem can be administered once daily (199). Although clinical st udies have demonstrated the effectiveness of ertapenem for SSSI-treatment, the in vivo penetration and the resu lting free protein-unbound concentrations in interstitial sp ace of soft tissues, such as skeletal muscle and subcutaneous 3 Copyright Oxford Journals, [J Antimicrob Chemother 58: 632-6, 2006]
71 adipose tissue have not been reported, mainly due to a lack of appropriate methodology (99, 151). One technique, which has been proven suit able for the measurement of target tissue concentrations of a variety of substances in vivo in humans, is clinical microdialysis (122). This method is a minimal invasive technique for the measurement of unbound drug concentrations in virtually every tissue and organ. The present microdialysis study was conducted to measure and compare the free protein-unbound ertapenem concentra tions in the intersti tial space fluid of two peripheral target sites, skeletal muscle and subcutaneous adipose tissue, following the administration of 1 g infusion, and to compare th em with the respective plasma concentrations. Volunteers and Methods Volunteers Six healthy volunteers (3 men, 3 woman, 4 Asians, 2 Caucasians), between 22 and 37 years old, average height 160.0 8.8 cm, average body weight 64.7 8.6 kg, average body mass index 25.3 2.4 kg/m2, and average body surface area 1.68 0.16 m2, participated in the study. All had normal renal and hepatic function; the mean creatinine clearance was 100.5 21.1 ml/min 1.73 m2. All volunteers included in the st udy had normal findings from physical examination, electrocardiogram and laboratory te sts (including haematol ogical and biochemical parameters, urinalysis, and negative pregnancy te st). The mean albumin serum concentration was 4.43 0.29 g/L, the mean total protein conc entration 7.50 0.24 g/L. Further exclusion criteria were regular use of medications, abuse of alcoholic beverages, symptoms of significant illness within 3 months before the study period, hi story of liver or kidney disease potentially interfering with metabolism or excr etion of the drug, history of cen tral nervous system disorders, allergy or hypersensitivity to the study drug, blood donation of more than 500 ml during the previous 3 months, participation in a clinical trial within 3 m onths before the study period and pregnancy. The study was conducted in the Gene ral Clinical Research Center at Shands
72 Hospital, University of Florida, was approved by Shands' Hospital Institutional Review Board (IRB-01), and was performed in accordance with the Declaration of Helsinki. All volunteers were given a detailed description of the st udy, and their written consent was obtained. Study Design and Protocol The study was conducted as a single-center, pros pective, open-label tria l. Volunteers were hospitalized from the evening befo re start of the study until 12 h post dosing. On the study day, the volunteers were kept under fasting conditions for 10 h prior to the start of the experiments until 2 h after drug administration. Each volunt eer received one dose of 1g ertapenem as an intravenous infusion over 30 min. Tolerability and safety assessments, clinical chemistry, haematological tests and urinalys is, and the measurement of vita l signs (blood pressure, heart rate) and ECG were included in th e study. Vital signs were take n before administration of the drug and 15 min, 1, 2, 12 h thereafter. All data relating to drug safety were recorded throughout the study. Each volunteer reco rded a diary protocol of possible adverse events. Sample Collection To measure the unbound fraction of ertapenem in the interstitial space fluids, microdialysis probes (CMA 60; CMA, Stockholm, Sweden) with a molecular cutoff of 20 kDa were used. Microdialysis probes were inserted after cleani ng and thorough disinfectio n of the skin. One dialysis probe was inserted into the medial vastus muscle and one was inserted into the subcutaneous layer of the thigh. The microdi alysis probes were secured in place by holding a plastic flap at the base of the probe while the needle was removed. The microdialysis system was flushed with lactated Ringer's solution and then connected to a microinfusion pump (Precidor; Infors-AG, Basel, Switzerland). The principles of microdialysis have previously been described in detail (27, 190, 192). Briefly, microdialysis is based on sampling of the nonprotein-bound fraction and, therefore, the pharmaco logically active fraction of analytes from the
73 interstitial space with a semiperm eable membrane at the tip of a microdialysis probe. The probe is constantly perfused with a phys iological solution (perfusate) at a flow rate of 2 L/min. Once the probe is implanted into the tissue, substances present in the extracellula r fluid at a particular concentration (Ctissue) are filtered out of the interstitial space fluid by perfusion into the probe, resulting in a concentration (Cdialysate) in the perfusate. Samples are collected and analyzed. For most analytes, equilibrium of the concentra tion between extracellular tissue fluid and the perfusion medium is incomplete; therefore, Ctissue > Cdialysate. The factor by which the concentrations are interrelated is termed recovery. To obtain absolute concentrations in the interstitial space fluid from the concentrations in unbound dialysat e, microdialysis probes were calibrated for in vivo recovery rates by the retrodialysis method (242). The retrodialysis procedure is performed in each subject before do sing of the drug. The principle of this method relies on the assumption that th e diffusion process is quantitativ ely equal in both directions through the semipermeable membrane. Therefore, ertapenem was added to the perfusate at a concentration of 5.0 mg/L, and the disappearance rate (deliv ery) through the membrane was taken as the in vivo recovery. The in vivo percent recovery was calculated as 100 ion Concentrat ion Concentrat 100) Recovery(%perfusate dialysate (4-1) After a 30-minute baseline perfusion period, in vivo calibration was performed as described previously for a period of 60 minutes, during which two samples were collected (at -30 and -60 minutes) (27). Concentrat ions in both retrodialysis sample s were averaged and used to calculate the in vivo recovery. After the ca libration period was completed, a 60-minute washout period was observed. Microdialysis sampling was performed at 60-minute intervals up to 12 hours post dose.
74 Blood samples (5 mL) were collected in lithium heparinate-coated tubes, via a venous plastic cannula (JELCO; Johnson-Johnson, Arlington, Tex), before ertapenem infusion, and 0.5, 1, 2, 3, 4, 6, 8, and 12 hours after the start of infusion. Samples were centrifuged at 1.300 g for 10 min at 4C. Plasma was separate d and stored at -80C until analysis. Drug Assay Quantitative determination of ertapenem in plasma and in interstitial space fluid of skeletal muscle and subcutaneous adipose tissue were determined by validated SPE-Liquid chromatographytandem mass spectrometry methods (LC-MS-MS) (132). The HPLC analytical column used was Synergi 4 Polar-RP 102.0 mm. The flow rate was 0.5 mL/min, with a gradient mobile phase of A = 2 mM ammonium acetate/0.1% Acetic Acid, pH = 3.2 and B = 100% methanol, starting from 100% A to 90% B in 5 minutes and change back to 100% A in 3 minutes, keep it there for 6 minutes, total of 14 minutes for each injection. MS-MS equipment and related parameters are: Micromass Quattro LC-Z, with negative electro spray ionization, MRM scan function for two channels: 474. 00 > 265.00 (ertapenem) and 466.00 > 422.00 (ceftazidime, as internal standard), cone vo ltage was 25 volts, collision energy was 20 for ertapenem and 10 for I.S.. The limit of quantification was 0.04 mg/L. Pharmacokinetic Analysis Pharmacokinetic parameters were calculated by non-compartmental an alysis with Kinetica pharmacokinetic software program (Kinetica 4.3, Innaphase). The maximum observed plasma concentration (Cmax) and time to reach Cmax (Tmax) after drug administration were determined directly from the concentration-time curves. The area under the plasma concentration curve from time 0 (the start of infusion) until the last quantifiable plasma concentration (AUC0-last) was calculated by using the log-linear trapezoidal rule. AUC0was derived by adding Clast/ z to AUC0-last. The terminal elimination rate constant ( z) was estimated from the slope of terminal
75 exponential phase of the logarithmic plasma con centration-time profile using no fewer than 3 data points. The apparent te rminal elimination half-life (t1/2z) was calculated as 0.693/ z. The mean residence time (MRT) was calculated AUMC0/AUC0, where AUMC0is the area under the first moment of concentration-time curve, determined by integr ating the product of time and concentration from zero to infi nity. Apparent total clearance (CLtot) was calculated as dose/AUC0. The apparent volume of distri bution during the terminal phase (Vz) was calculated as (CLtot/ z). Protein-unbound ertapenem concentrations in the extracellular skeletal muscle and subcutaneous adipose tissue fluid were calculated from measured microdialysate concentrations and individual probe rec overy, determined in our in vivo experiments. The parameters, such as Cmax, Tmax, AUC0-last, AUC0, and t1/2z were calculated using the same formula as for plasma samples. The tissue penetration was calculated as the ratio of the unbound AUC0in skeletal muscle or subcutaneous adipos e tissue fluid to the total AUC0in plasma (AUCtissue, free/AUCplasma, total). All data are presented as geometric means standard deviations, with the exception of Tmax, for which median and minimum-maximum ranges are given only. Results Safety All six enrolled volunteers completed the study in accordance with the protocol. The microdialysis procedure and treatments were well to lerated. No serious or severe adverse events or microdialysis-associated side-effects were observed. Two volunteers reported headache, not related to study drug. There were no clinically significant cha nges in electrocardiograms, blood pressure or pulse. Similarly, there were no clin ically important findings in haematology, clinical chemistry or urinalysis.
76 Pharmacokinetics The time versus concentration profiles of ertape nem in plasma (total concentration) and in the interstitial space fluid of sk eletal muscle and subcutaneous adipose tissue after administration of a single intravenous dose of 1 g over 30 min to healthy volunteers (n = 6) are shown in Figure 4-1. Pharmacokinetic parameters are listed in Table 4-1. Discussion Bacterial skin and skin-structure infections are among the most frequently seen infectious diseases in the community and occasionally in th e hospital setting (32, 71 ). Larger and profound lesions are usually secondarily mixed-infected with aerobic Gram-positive and Gram-negative bacteria including Staphylococcus aureus, Streptococcus species, Enterobacteriaceae, and anaerobic bacterial pathogens (71, 92, 96). Ap art from surgical and general care measures (wound cleaning) a therapy with highly effective antimicrobial agents is indicated, because infected SSSI are often starting point of phlegmonous inflammations and in worst case also of a septic syndrome (32, 71, 92, 96). For the selectio n of a suitable antibiotic drug the antibacterial spectrum and the concentrations r eached at the site of infection are very important (152, 188). Many studies have shown that plasma concentrat ions may not be an ideal parameter for the prediction of the clinical efficacy of antibiotics b ecause most infections occur at the tissue sites. Additionally, it has been found that only free protein-unbound an tibiotic concentrations at the infection sites are responsible for the antibacterial ac tivity (136, 177). Therefore, inadequate interstitial tissue concen trations can lead to therapeutic fa ilure and bacterial resistance (152, 188). Due to its wide range of antibacterial activity against most Gram-positive, Gram-negative and anaerobe bacteria, ertapenem is a candidate for the treatment of skin and skin-structure infections (56, 156, 281). However, there is only limited information on the ability of ertapenem
77 to penetrate into different organ tissues, such as lung or pancreatic ti ssue (35, 273). With reference to the treatment of SSSI, one study investigated ertapenem penetration into suctioninduced skin blister fluids in 12 healthy young volunteers (139). Drug concentrations in skin blister fluids exceeded 4 mg/L (the MIC at wh ich 90% of isolates tested are eliminated) throughout the entire dosing interval of 24 hours. However, extrap olation of these data to the concentrations in infected tissues should be done with extreme caution (27). One problem is that skin blisters are formed before the administration of the antibiotic. The blister thus serves as a large third compartment with a surface-to-volume ratio, which is hardly representative of that for tissue. Besides, other variables must be take n into account, such as the barrier between the blister and the skin, which may change over time and the presence of protei ns in blister fluid. The main purpose for our study was to measur e the concentrations of the non-proteinbound ertapenem in interstitial fluids of two di fferent SSSI target sites (skeletal muscle and subcutaneous adipose tissue) by microdialysis after administration of a single intravenous standard dose (1g/day). The results show that ertapenem reaches measurable concentrations in both target tissues. As expected time versus concentration profiles indicate that free ertapenem concentrations in interstitial space fluids were lower than the corresponding total concentrations in plasma (Figure 4-1). The tissue levels meas ured in our study corresponded approximately to the free protein-unbound fraction of ertapenem in plasma (4-16%). Free interstitium / total plasma concentrations ratios for skeletal musc le and subcutaneous adipose tissue were not congruent, a finding that has been observed previously and that might be explained by differences in local blood flow between the two tissues (27). Finally, the question arises: Ar e the free, protein-unbound ertapenem concentrations in the interstitial space fluids of muscle and subc utaneous adipose tissue high enough to kill the
78 bacteria effectively? Like other -lactam antibiotics, carbapenems exert their killing effect in a time-dependent manner (53, 231). In this category of antibacteria l drugs, increasing the concentrations above 4 5 times the MIC of the bacteria no longer adds a proportional increase in the killing effect. Therefore, maximum killin g is obtained by optimizing the time of exposure of the drug to the bacteria so that the concentrat ions remain above the MIC as long as possible. The main pharmacokinetic/pharmacodynamic (PK/PD) parameter for -lactams is the proportion of time of the dose interval during whic h the drug concentration exceeds the MIC (T>MIC). For carbapenems a T>MIC of 30 to 40% of the dose interval ha s been previously suggested to be effective due to their rapid bactericidal activity (186). In vitro studies demonstrated that ertapenem inhibited 90% (MIC90) of methicillin-susceptible Staphylococcus aureus strains at 0.25 mg/L (91, 154). Against Streptococcus spp., ertapenem had a MIC of 0.5 mg/L and against extended-spectrum -lactamase (ESBL)-producing Enterobacteriaceae MIC90 values ranged from 0.03-0.06 mg/L (91, 154, 155). Ertapenem MIC90 values for Bacteroides fragilis and other anaerobe bacteria were 1.0 mg/L (154). In this study, protein-unbound ertapenem concentrations 12 h after a single intravenously ad ministration of 1 g were 1.13 0.68 mg/L for muscle tissue and 0.31 0.16 mg/L for subcutaneous adipose tissue. Therefore, a 1 g dose once daily results in muscle tissue concentrations higher than MICs of most above-mentioned SSSI pathogens for at least 50% of the entire dosing interval. Also the mean concentrations in subcutaneous adipose tissue exceeded the time above the MIC90 of SSSI pathogens by at least 30% (Figure 4-1). From this finding we can conclude, that the data obtained in this study suggest adequate free, protein-unbound ertapenem concentrations in th e non infected interstiti al fluid of muscle
79 and subcutaneous adipose tissue. Our results su pport the previously observed clinical efficacy of ertapenem in the treatment of skin and skin-structure infections.
80 Table 4-1. Non-compartmental pharmacokinetic analysis of ertapenem after 1 g single intravenous dose Parameters Plasma Muscle Subcutis Cmax (mg/L) 103.3 26.3 6.71 4.14 3.96 1.63 Tmax (h) 0.5 (NA)2.0 (1.0 3.0)1.0 (1.0 2.0) C12 (mg/L) 7.93 4.15 1.13 0.68 0.31 0.16 AUC0-last (mgh/L) 316.1 49.1 36.7 23.4 17.4 4.0 AUC0(mgh/L) 359.7 66.5 41.1 26.0 19.5 4.9 Terminal half-life (h) 3.77 0.60 3.38 0.68 3.63 0.85 MRT (h) 4.58 0.88 ND ND Vz (L) 15.5 3.4 ND ND Cltot (L/h) 2.88 0.51 ND ND AUCinterstitiu m /AUCtotal p lasma 0.13 0.10 0.05 0.01 Data are presented as geometric mean sta ndard deviation, with the exception of Tmax, for which median and minimum-maximum ranges are given. C12 = Ertapenem concentrations obtained 12 h after start of infusion.
81 Figure 4-1. Comparison of ertapenem concentration pr ofiles in plasma (total concentration) with unbound tissue concentrations in skeletal musc le fluid and intersti tial adipose tissue fluid in healthy volunteers (n = 6) after a single intravenous do se of 1 g (infusion period 30 min). Horizont al lines indicate MIC90 values for Bacteroides fragilis (), Streptococcus spp. ( ), methicillin-susceptible Staphylococcus aureus ( ), and extended-spectrum -lactamase (ESBL)-produc ing Enterobacteriaceae ().
82 CHAPTER 5 INTEGRATION OF MODELING AND SIMULA TION IN DEVEL OPMENT OF NEW ANTIINFECTIVE AGENTS MIC VS. TIME-KILL CURVES4 Introduction Historically, clinical pharmacology is divi ded into two main areas pharmacokinetics (PK) and pharmacodynamics (PD). While PK de scribes what the body does to the drug, PD evaluates what the drug does to th e body. In isolation, PK and PD are of limited usefulness and a mechanism-based linkage of PK and PD is essential in order to design a rational drug development plan (68, 69, 175, 176). In practice, numerous complex interactions between drug, patient and disease state can make this task quite challenging. Identification of specific, consistent and reliable biomarkers offers a so lution to this challenge and is commonly employed in rational drug development (47). Once suitable biomarkers are identified, they are incorporated into PK/PD modeling and simulatio n approaches in order to predict clinical endpoints of new doses and dosing regimens. In the area of infectious diseases, to date, minimum inhibitory concentration (MIC) has fu lfilled this role as a well-established and routinely determined PD paramete r for antibiotics. According to the Clinical and Laboratory Standards Institute (CLSI, formerly National Committee for Clinical Laboratory Standards NCCLS), MIC is defined as the lowest concentration of drug that completely inhibits visible growth of the organism as detected by the unaided eye after an 18-24 hour incubation period with a standard inoculum of approximately 105 CFU/mL (188). MIC-ba sed PK/PD indices such as T>MIC, AUC/MIC and Cmax/MIC have been introduced into anti-infective drug therapy (182, 183). Although this use of MIC combined with pharmacokinetic parameters has led to a much better understanding of antibiotic do sing, the MIC is still an impr ecise threshold with several 4 Copyright Informa Pharmaceutical Science, [Expert Opin. Drug Discov. 2: 849-860, 2007]
83 limitations (e.g. inter MIC test variability, inab ility to account for changes in free antibiotic concentration over time) that does not allow prediction of antimicrobi al activity at concentrations apart from the MIC itself (148, 216, 231). The purpose of this review paper is to 1) evaluate MIC and MIC-based PK/PD indices, 2) introduce an alternative PK/PD mo deling approach, 3) discuss stre ngths and weaknesses of both approaches and 4) give a perspective for future research in the field. PK/PD Strategies for Anti-Infectives Effect of Drug Binding Most drugs can be bound to proteins or other biological material and hence are present in the body in bound and free form. It is a well-recogni zed fact that for small molecules only free drug is pharmacologically active (13, 59, 200). Th erefore, for PK/PD modeling, the use of free drug is recommended and indicated by the prefix in the PK input function (183). In most cases, free, unbound drug concentrations in well-per fused tissue seem to correlate well with unbound concentrations in plasma and can ther efore be estimated from total plasma concentrations (59). Since alteration of the fraction unbound (fu) does not necessarily result in changes in free steady-state conc entrations (45, 255), oversim plifications estimating free, unbound concentrations especially of highly bound dr ugs might lead to wrong conclusions. This scenario becomes even more complex in pa tient populations where altered physiological conditions such as liver function, uremia or hypoalbuminemia need to be taken into consideration. It has further been show n that protein binding results determined in vitro might need to be interpreted with care (200). Ther efore, a close evaluation including metabolic and disease status is necessary. The actual measur ement of free, unbound concentrations at the site of action/infection using microdialysis is an alte rnative way to precisely determine the PK input function.
84 MIC-Based PK/PD Indices MIC-based indices are intended to normali ze the drug exposure (PK) to the respective sensitivity (MIC). The first PK/PD index was developed for penicillins where in vivo antimicrobial efficacy was correlated to the tim e free or total drug levels stay above the MIC (T>MIC or T>MIC, respectively) of the target organism. Other commonly used indices are the free or total maximum plasma con centration over MIC ratio (Cmax/MIC or Cmax/MIC) and free or total area under 24 hour plasma concen tration time curve over MIC ratio (AUC24/MIC or AUC24/MIC) (147, 183, 225). Although some author s suggest a further index differentiation within the drug class (6, 52, 272), T>MIC 40% of the dosing interval seems to be sufficient for clinical efficacy of -lactam antibiotics (49-51, 90, 167, 170, 237). In comparison, Cmax/MIC values of 10-12 seem to be a good predictor for aminoglycosides (125, 180, 223, 256). For fluoroquinolones, AUC24/MIC or Cmax/MIC ratios seem to be the best index (8, 113, 201, 205, 223, 272). Target AUC24/MIC values of 100-125 have been identified for gram-negative bacteria (39, 49, 89, 252), 25-35 for gram-positives pathogens (49) and Cmax/MIC values of 10 (205). The magnitude of the fluoroquinolone PK /PD index needed is still controversially discussed and seems to vary with type of infection, fluoroquinolone, pathogen, immune status and degree of protein binding (6, 205). At this point, we strongly agree with other authors stating that only free, unbound and pharmacologi cally active fraction should be used and appropriately marked to calculate respective PK/PD indices (4, 183, 272). Based on these MIC-based PK/PD indices dosing, regimens can be optimized to ensure efficacy and minimize the emergence of resistance This becomes particularly important in special patient populations, such as neonates or elde rly, where there is a lack of sufficient data on the efficacy of dosing regimens (62). In drug development, the good correlation of clinical
85 outcome with T>MIC, Cmax/MIC and AUC/MIC makes them a valuable predictor for the efficacy of new drugs or dosing regimens. For example, treatment of drug-resistant pulmonary tuberculosis (TB) with oflox acin 10mg/kg QD was successfully suggested to be sufficient based on determined Cmax/MIC ratios (7.7-15.4) and a long half-life (46). However, relationships based on MICs su ffer from numerous drawbacks. While antibacterial activity is a dynami c process, MIC is a crude mono-dimensional threshold value that shows a high variability due to its determ ination (202). Depending on the method used (Etest, Kirby Bauer test, broth/agar dilution method, etc.), results can vary dramatically. As a consequence, MICs are often reported as a range of concentrations instead of a single value. Consequently, MIC-based indices and predictions are highly variable as well. Hence, it was suggested that for any calculati on or expression of the MIC a desc ription of the method by which the MIC was determined should be mentioned (182, 183). At present, re gulatory authorities are also working on global standard ization of MIC methods (183). Time-Kill Curve-Based PK/PD Indices To overcome limitations of the MIC, s ubsequent sampling of bacterial counts in vitro or in animals can be used to provide a time course of antimicrobial action (time-kill curve). In general, there are two different types of time-kill curves, based on constant concentrations, or on fluctuating concentrations (dynamic models) (226 230, 231). The constant concentration model simulates the steady-state concentration obtained in vivo after constant rate infusion. In this approach a set of culture flas ks containing bacteria and nutri ent are exposed to different antibiotic concentrations, usually multiples of the MIC, for a certain incubation time. Samples are taken at preset time points a nd plated on appropriate agar plates. Bacterial counts are taken and then plotted as CFU/mL against time.
86 Dynamic models try to reflect the change in drug concentration that occurs in vivo (Figure 5-1) (164). In order to simulate the half-life, drugcontaining medium is removed from the flask and replaced by fresh medium. This can either be done manually using syringes or automatically by employing pumps. Microfilters (0.2 m) placed between flaskand pump system prevent the bacteria from washout. Additional flasks can be used to simulate a multi-compartment scenario. This ability to mimic the PK of a drug for oneand multi-compartment-body models makes the dynamic model a very powerful PK/PD prediction tool (231). Detailed descriptions of the experimental setting of these models and thei r applications can be found elsewhere (164, 188, 198, 206). Once the time-kill curve experiments are performed, a mathematical model can be identified to describe the data. Before discussing these models in greater detail, a general concept of how they link PK and PD together is provided in th e following section. The rationale for PK/PD-modeling is to link PK and PD in order to establish and evaluate dose-concentration-response relati onships and subsequently descri be and predict the effect-time courses resulting from certain doses (69, 175). Keeping in mind that all models are wrong, but some are useful, PK/PD models represent a simplif ication of the real physiological process (22). They can either be purely descriptive, regardless of the actual mechanism of action, or they can be mechanism-based appreciating the underlying phys iological events (69). In any case, PD effects can either be directly correlated to plasma concentrations (direct link model) or show a time delay (indirect link model). Plotting effect vs. plasma concentration results in a single curve for direct link or a hystere sis loop for indirect link models, respectively (38). Definition of a hypothetical effect-compartment (Figure 5-2) offers an elegan t solution for this problem and leads to a collapse of the hystere sis loop in the concentration-eff ect relationship (69, 109, 110). Hence, either the measured or calculated free plasma concentration (direct link model) or free
87 concentration in the effect compartment Ce (indirect link model) serves as the PK input function. This PK input function is then combin ed with an appropriate PD model. However, drug interacting with its target at the effect site may not necessar ily result in a direct response. Whenever, the mechanism of action involves a ti me-dependent physiological process (e.g. the down-regulation or synthesis of re ceptors or transporters), the re sponse will occur with a certain time delay. In such cases, indirect response mo dels are fitted to the data accounting for the temporal dissociation between concentrationand effect-time courses. Pharmacodynamic Models To appropriately describe effects over time relationships, different PK/PD models have been established (67). Of those, Emaxand sigmoidal Emax-models are most commonly used in PK/PD modeling. Especially modified sigmoidal Emax-models have found great acceptance in describing PK/PD of antibiotics. An example is shown in equation 5-1 N CEC Ck k dt dNh e h 50 h e max 0 f f (5-1) where dN/dt represents the change in number of viable bacteria, usually given as colony forming units per millilitre (CFU/mL) over time (188, 258). Growth in the absence of antibiotic is characterized by the bacter ial growth rate constant (k0) reflecting the rate of multiplication of a given pathogen in its surrounding environment. Ho wever, in the presence of antibiotic, bacterial kill can be described by an Emax-model with the maximum killing rate constant (kmax), the free concentration of antibiotic necessary to produce 50% of the maximum effect (EC50) and the free concentration of antibiotic in the effect compartment (Ce) at any time (t). The general curve fit can be further optimized with a Hill or shape fact or h that modifies the steepness of the curve. A delayed onset in growth and/or kill might make the addition of further factors necessary (258).
88 A brief outline of the mathematical relati onships underlying time-kill curve modeling was provided by Mueller et al. and is illustrated in th e following scenario (188). Starting out with an initial inoculum N0 of 5x105CFU/mL, a hypothetical bacterial stra in X grows in the absence of antibiotic (Ce = 0) with k0. Plotting the number of bacteria at time t (Nt) vs. time on a semi-log scale results in a straight lin e with a growth rate of k0 as the slope as shown in equation 5-2 and Figure 5-3. 0 0 tk t lnNlnN slope (5-2) However, it should be noted that Figure 5-3 s hows ideal growth behavior. In practice, the in vitro exponential growth phase is limited by nutriti onal and environmental constrains resulting in slower growth rates that can be modelled by additional parameters (258). These factors are dependent on strain(s), antibiotic (s) and experimental conditions used (172, 194, 258, 264). For didactic reasons, these factors were not include d in the simulations presented in Figure 5-3. Addition of an antibiotic A to the inoculum results in a diminished net-growth. From a certain drug-specific Ce, a maximum effect will be reach ed, where a further increase in Ce does not contribute to the antimicrobial effi cacy any longer. Solving equation 5-1 for Ce >> EC50 leads to equation 5-3. N C Ck k dt dNh e h e max 0 f f (5-3) At such high antibiotic concentrations the antimicrobial effect is not dependent on Ce any longer. The slope of this maxim um effect curve is defined by equation 5-4 as the difference of k0 and kmax (Figure 5-3).
89 Nkk dt dNmax 0 (5-4) Since EC50 is defined as the concentration of an tibiotic necessary to produce 50% of the maximum effect kmax, it can be illustrated as the bisecting line of the angle formed by growth control (zero effect) and maximum kill effect as shown in Figure 5-3. A general statement whether the EC50 is equal to, smaller or bigger than the MIC cannot be made. A special case occurs when antibiotic induced kill equals growth and so the number of bacteria stays constant resulting in a horizontal line. The respective concentration of an tibiotic in this scenario is referred to as stationary concentr ation (SC) (187). It can be calc ulated using equation 5-5 and is different than the MIC. 50 h 1 0 max 0EC kk k SC (5-5) However, to date, MIC is more commonly us ed as a pharmacodynamic measure than SC. Being defined as the lowest concen tration of drug that completely inhibits visible growth of the organism as detected by the unaided eye after an 18-24 hour incubation period, the MIC value is linked to a turbidity threshold. This threshold is reached for most bacteria at concentrations of approximately 107 CFU/mL (Figure 5-3) (188, 202). Substitution of Ce by the MIC in equation 5-1 leads to equation 5-6. N MICEC MICk k dt dNh h 50 h max 0 (5-6) Rearranging and integration of equation 56 on both sides results in equation 5-7.
90 t 0N N t 0 h h 50 h max 0dt MICEC MICk kdN N 1 (5-7) As shown in Figure 5-3, the MIC turbidity limit is reached after an incubation time t of 18 hours. Solving the integral in eq uation 5-7 with a lower limit N0 of 5.5x105CFU/mL and an upper limit NMIC of 107CFU/mL leads to equation 5-8. h h 50 h max 0 0 MICMICEC MICk k t lnN lnN (5-8) Rearranging for the MIC leads to equation 5-9. 50 h 1 0 MIC 0 max 0 MIC 0EC N N ln t 1 kk N N ln t 1 k MIC (5-9) Considering that the concentr ation of antibiotic remains c onstant throughout the entire MIC experiment, kmax, EC50 and h remain the only variables in equation 5-9. Since k0 is a constant for a given strain X in a given medium as well, k0, N0, NMIC and t can be combined to a drug independent constant d as shown in equation 5-10 (188). 50 h 1 maxEC dk d MIC (5-10) The resulting MIC equation has three unknow ns and hence it cannot unambiguously describe the effect-time course for a given antibiotic concentration. Consequently, kmax, EC50 and h have to be determined separately to clear ly specify the antimicrobial properties of a drug
91 against a certain pathogen. Equa tion 5-10 clearly illustrates that the same MIC value can be a product of different combinations of kmax, EC50 and h. Once these relationships have been evaluated and validated, time-kill curve based indices can be used to compare different doses, dosing regimens or dosage forms. The suitability of a new ciprofloxacin extend ed-release dosage form (1000mg, QD) was compared to the outcome of the standard dose (500mg, BID) (230). Ciproflo xacin PK parameters were determined from mean serum concentrations (N=19) using non-co mpartmental and compartmental data analysis and linked to PD parameters determined from in vitro time-kill curve experiments against E. coli. Comparison of expected kill curves with th e immediate-release versus extended-release treatments showed similar outcome. Therefore, a once daily dosing regi men was suggested, due to its similar efficacy but superior patient complia nce compared to traditional immediate release. PK/PD Simulations Monte Carlo simulation Both MIC and time-kill curve parameter do not a ccount for variability in either patient or bacterial population. Instead, mean values or concentration range s are evaluated. One approach to incorporate variability among clinical PK a nd PD is using Monte Carlo simulation (MCS) as an advanced statistical modeli ng tool. MCS allows prediction of therapeutic outcome based on an integrated PK/PD stochastic model for a drug or dosing regimen (2, 197). The first step in a MCS is to estimate mean PK parameters and thei r associated variability for a desired patient population and dosing regimen. This can be done by using analysis tools such as population pharmacokinetics (POPPK) (231). From this mu ltivariate distribution data pool, values are randomly picked and a PK profile is simulated. This random selection process is repeated multiple (e.g. 5,000) times and respective PK prof iles for these hypothetical patients are obtained (159). Together, they give a range of possible PK profiles. This information can now be used to
92 estimate the probability with which a predefined PK/PD index (e.g. T>MIC = 40%) is reached for that particular MIC. Whether this target is attained for a range of MICs can be calculated by summing up the products of frequency of specific MI C values and their respective probability of target attainment (159). This ab ility of predicting the likelihood of a certain outcome to occur has been used by numerous investigators to optim ise dosing of drugs or drug combinations in both preand post-marketing st udies (137, 159, 184, 231). This approach has also been applied to compare potencies of different drugs, determ ine susceptibility break points for initial human trials and help selecting drugs (5), doses and dosing regimens (65, 77, 138, 160) in healthy, renal impaired (248) and elderly pa tient populations (3, 160, 203). In a recent study, PK modeling and MCS (10,000 s ubjects) were performed to describe the PD profile of cefepime in plasma and cerebrospinal fluid (CSF) (158). PK data of 7 hospitalised patients with external ventricular drains was used for MCS in order to es timate the probability of attaining targets of cefepimeplasma (20% plasma protein bi nding) and total cefepimeCSF of 50 to 100%T>MIC for MICs of 0.06-8mg/L. Calculations we re performed for short-term infusions (0.5 hours) of 2g cefepime TID and BID or QD plus c ontinuous infusion (250mg/h). Results indicate that in plasma, targets attainment rates of T>MIC 60-70% were high at each MIC (0.03-8mg/L) for all dosing regimens examined. Howe ver, at MICs > 0.5mg/L, indices of T>MIC 50-100% were not reached for more than 80% of the patie nts. As shown, MCS is an important simulation tool employed in the establishment of new dosing regimens. To date, most MCS are performed using MIC rather than model based PK/PD indices as the therapeutic target. Since the MIC is not an intrinsic property of a microorganism, its use as a single point estimate oversimplifie s the dynamic process of growth and natural or antibiotic
93 induced kill (187, 226). In future approaches the use of time-kill parameters in MCS approaches might provide more detailed information than the MIC. Time-kill curve based simulations The relationships derived in the previous sectio ns are not only of theoretical interest but can also be used to simulate clinical outcome. In the following it will be shown that time-kill curve based indices provide more reliab le information than MIC based indices. Simulations were performed for the -lactam antibiotics faropenem and ceftriaxone for a TID and BID dosing regimen, respectively. It has been shown for -lactams that the free antibiotic concentration has to exceed the MIC for at least 40% of the dosi ng interval in order to be clinically effective. Maximal efficacy for ce phalosporins in several animal infection models was shown for Enterobacteriaceae and Streptococci when T>MIC in serum exceeded 60%-70% of the dosing interval (51). Therefore, br eakpoint values of T>MIC = 40% and T>MIC = 67% (2/3 of the dosing interval) have been chosen fo r our simulation. Intraven ous doses necessary to maintain steady state concentrations above these breakpoints were calculate d from PK data given in the literature, using a one-compartment body mode l with a first order elimination rate constant ke and correcting for plasma protein binding (1 30, 244). Mean plasma protein binding for ceftriaxone and faropenem was assumed to be 90% (211, 212) and 96% (20), respectively. Respective PD parameters of Streptococcus pneumoniae CDC 145, Streptococcus pneumoniae ATCC 6303 and Haemophilus influenzae ATCC 10211 were determined and are listed in Table 5-1 (129, 130, 153). In the first scenario (A), k ill-curves of faropenem against Haemophilus influenzae ATCC 10211 and ceftriaxone against Streptococcus pneumoniae CDC 145 were simulated and the outcomes were compared. For T>MIC = 40%, failure in antimicrobial treatment is predicted for both drugs (Figure 5-4 A). In comparison, treatment with faropenem is able to reduce the
94 number of H. influenzae by ~ 3 log steps (Figure 5-4 B), whil e treatment with ceftriaxone still remained ineffective against S. pneumoniae for T>MIC = 67%. Faropenem given TID appeared to be more effective against H. influenzae ATCC 10211 than ceftria xone given BID against S. pneumoniae CDC 145. However, a direct comparison of both dosing regimens is rather difficult, since S. pneumoniae CDC 145 and H. influenzae ATCC 10211 show diffe rent growth rates (equation 5-9, Table 5-1). Hence, a second set of simulations (B) was pe rformed evaluating the antimicrobial potency of faropenem (TID) and ceftriaxone (BID) against S. pneumoniae ATCC 6303 for T>MIC = 40% and T>MIC = 67%, respectively. Simultaneous simulations of scenario B predicted a failure of the treatment with ceftriaxone for both dosing regimens (Figure 5-4 C and 5-4 D). The T>MIC = 40% dosing regimen is predicted to be ineffec tive for faropenem, while total eradication of S. pneumoniae ATCC 6303 is predicted afte r 36 hours, adjusting for T>MIC = 67% (Figure 5-4 D). Based on T>MIC = 40%, a false positive treatment outcome is predicted for both treatment regimens. Discussion and Conclusion In comparison to the single endpoint MIC, modelled time-kill curve parameters provide more detailed information. These parameters desc ribe the antimicrobial e ffect over time. This additional information can then be further used to optimise the dosing regimen. However, timekill curve analysis is a very labour intensive method. Due to the fact that quite a large number of time-kill curves are necessary to allow sufficient evaluation of concentrationand time-effect relationships, time and manpower are limiting fact ors. Considering the current speed of technical development, these limitations ar e likely to be counterbalanced by improved experimental analysis technique s and automation in the future.
95 Since MIC and time-kill curves are performed in vitro, both have limitations with respect to their predictability of in vivo situations. Conditions simulated in vitro are much less complex than the actual in vivo situation. Drug can be bound to proteins, growth-rates might be altered in human fluids and infection is usuall y affected by the immune system. In vivo protein binding and altered growth can be simulated by performi ng time-kill curves in human serum or animal experiments (75). In many cases, the immune system of the test animal has to be knocked out in order to facilitate an infection. However, a complete suppression of the host defense may only apply to a special patient popul ation undergoing chemotherapy (e .g. HIV, cancer, transplants). Therefore, an extrapolation of results determined in vitro to in vivo can lead to an overestimation of the antibiotic effect (231). Although these experiments cannot fully reflect the situation in humans, valuable data can be generated for th e development of dosing regimens and therefore have become inexpensive alternatives to clin ical trials (143, 268). However, the above discussion assumes a homogenous population of bacteria with a single MIC or EC50, respectively. In reality this is rarely the case and more complex approaches are needed to adequately evaluate these situations (37, 75, 146, 230). Based on data from in vitro, animal experiments and clinical trials, mathematical models can be developed that provide insight into th e population dynamics underl ying the emergence of antimicrobial resistance. Models accurately pr edicting the outcome of antimicrobial therapy can then be used to evaluate different dosing re gimens that are able to prevent resistance development (37, 100). While emergence of resi stance is indicated by increasing MIC values, only with continuous sampling over time, distinct evaluation of resistance development can be assessed. However, models needed to describe these resistance scenarios can become fairly complex and much more work needs to be done to develop and validat e new time-kill curve
96 based resistance models. Once these models have been established and validated, much more information can be extracted from them than from MIC indices (230). In conclusion, both MIC and time-kill curv e approaches have limitations but their introductions into antimicrobial therapy were milestones in making educated dose recommendations compared to previous empiri cal treatment of most likely disease causing pathogens.
97 Table 5-1. Pharmacokinetic and pharmacodynamic parameters for ceftriaxone and faropenem used for the simulations. Parameter Scenario A Scenario B Faropenem versus H. influenzae ATCC 10211 Ceftriaxone versus S. pneumoniae CDC 145 Faropenem versus S. pneumoniae ATCC 6303 Ceftriaxone versus S. pneumoniae ATCC 6303 MIC determined (mg/L) 0.50.64-1.280.01-0.02 0.01-0.02 MIC calculated (mg/L) 0.42 1.09 0.01 0.018 Dosing interval (h) 8 12 8 12 Vd (L) 15.9 7 15.9 7 ke (h-1) 0.693 0.087 0.693 0.087 fu in plasma 0.04 0.1 0.04 0.1 k0 (h-1) 0.98 1.44 1.67 1.5 kmax (h-1) 1.79 2.47 3 2.6 EC50 (mg/L) 0.5 1.063 3 2.6 h 1 2.5 1 2.5
98 Figure 5-1. Design of in vitro model. (A) Culture vessel; (B) construction to clamp the upper and bottom part together; (C) magnetic stirrer; (D ) side arm for sampling; (E) side arm for supplying fresh medium; (F) vessel containing fresh medium; (G) pump.
99 Figure 5-2. Two-compartment body model with additional effect compartment
100 Figure 5-3. Growth/kill curves at differ ent drug concentrations to illustrate k0, kmax, EC50, MIC and SC relationship
101 Figure 5-4. Simulated time-kill curves and respective growth controls of faropenem against Haemophilus influenzae ATCC10211 (solid line) and ceftriaxone against Streptococcus pneumoniae CDC145 (dotted line). For th e therapeutic target A) T>MIC = 40% and B) T>MIC = 67%, as well as, faropenem (dash-dotted line) and ceftriaxone against Streptococcus pneumoniae for the therapeutic target C) T>MIC = 40% and D) T>MIC = 67% compared to grow th control (solid line).
102 CHAPTER 6 PHARMACOKINETIC/PHARMACODYNAMIC MODELING OF THE IN VITRO ACTIVITY OF OXAZOLIDINONE ANTIBIOTICS AGAINST METHICILLIN-RESISTANT STAPHYLOCOCCUS AUREUS5 Introduction Gram-positive pathogens are a major cause of a wide range of infec tions, including skin and skin structure infections (SSSI) and life-threatening infecti ons like bacteremia, endocartitis and pneumonia (163, 227). Treatment of these infections has become increasingly challenging due to the rapid development of resistance against first choice antibiotics. In many cases, physicians are left with no other choice than usin g highly potent last resort antibiotics, such as vancomycin, on a daily basis. However, with the appearance of vancomycin-resistant strains, for example vancomycin-resistant Staphylococcus aureus (VRSA) or vancomycin-resistant enterococci (VRE), the once powerful antibiotic arsenal has become ineffective (42, 124). Therefore, it is critically impor tant to develop new potent antibi otics or antibiotic classes to successfully treat these infections. In 2000 the FDA approved with linezolid the first representative of a novel class of antibiotics, the oxazolidinones. At the time of approval, linezolid was one of the few agents that showed activity against vancomycin-resistant stra ins (40, 72). However, resistances against linezolid have been reported as early as 2002 (10 4, 259). In addition, toxic side effects such as reversible thrombocytopenia, neutropenia or rarely neuropathy have occurred during prolonged use (43, 149, 239). Due to these limitations ther e is a definite opportunity to develop new oxazolidinones with improved pharmacokinetic/pharmacodynamic (PK/PD) properties. RWJ416457 is a new investigational oxazo lidinone that is being devel oped as both an oral and an intravenous formulation for the treatment of infections caused by clinically important Gram5 Manuscript under submission with Antimicrob Agents Chemother.
103 positive bacteria. Compared to linezolid, RWJ416457 has a Minimum Inhibitory Concentration (MIC) that is twoto four-f old higher antimicrobial activity against multidrug resistant Grampositive pathogenic bacteria, including methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-intermediate susceptible Staphylococcus aureus (VISA), VRSA, VRE, and penicillin-resistant streptococci (87, 157). Although the MIC is r outinely determined in clinical settings and has contributed much to the unders tanding of antibiotic dosing, it does not provide any information on the time course of bacteria l growth or antibiotic-induced kill (188, 231). More detailed information can be obtained from the evaluation of growth and kill profiles over time (time-kill curve). A major strength of the time-kill curve approach is its capability of simulating the effect of changing concentratio ns on the antimicrobial outcome. Changing concentration time-kill curves can, subsequently, be used to evaluate the efficacy of antibiotics with different half-lives (t1/2). Once these experiments have been performed, a mathematical model can be simultaneously fitted to the data and respective PD parameters calculated. These PD parameters can then be linked to in vivo PK information to predict clinical outcome. The aim of this study was to 1) establish a ge neral mathematical model that is appropriate for characterizing the in vitro PD of oxazolidinones determined in constant, as well as, changing concentration time-kill curve experiments and 2) to apply this model in order to compare the in vitro potencies of investigational RWJ-416457 (t1/2 ~ 24h) and first-in-class representative, linezolid (t1/2 ~ 5h). Material and Methods Antibiotics and Growth Media RWJ-416457 and linezolid were provided by Jo hnson & Johnson Pharmaceutical Research & Development L.L.C. (Raritan, NJ, USA). Compounds were stored at 4oC in the original opaque vials. RWJ-416457 and linezo lid stocks were prepared fresh daily prior to use, kept at
104 room temperature and diluted to the desired c oncentrations with Mueller-Hinton broth (MHB; DIFCO, Lawrence, KS, USA). MHB was prepared according to the manufacturers manual and autoclaved prior to use at 121oC (15 min per 1L). Organisms MRSA OC2878 was obtained from Johnson & Johnson Pharmaceutical Research & Development L.L.C. (Raritan, NJ, USA). To ensure purity of the MRSA strain, colonies were plated on 5% sheep blood agar plates (Remel Microbiology Products, Lenexa, KS, USA) at least three times from overnight cultures before usage. Bacterial inocula for MIC and time-kill curve ex periments were prepared in sterile saline solution and adjusted with MHB to a fi nal concentration of approximately 5x105 colony forming units per milliliter (CFU/mL). MIC Determination MIC values of RWJ-416457 and linezolid agai nst MRSA OC2878 were determined in 24well plates (Corning Inc., NY, USA) using a modified Macrodilution Broth method (258). The first dilution where no visual turbidity appeared after a 20-hour incubation period was determined to be the MIC. The procedure wa s repeated six times pe r bacterial strain and antibiotic. Positive controls (with bacteria, no drug) and negative controls (no bacteria, no drug) were run simultaneously in order to assess the method. Constant Concentration Time-Kill Curves An in vitro model was used to investigate the e ffect of constant RWJ-416457 and linezolid concentrations on MRSA OC2878. This in vitro model consisted of eight 50mL cell culture flasks (NuncTM, Nunc A/S, Roskilde, Denmark). Flas ks were filled with 20mL MRSA OC2878containing MHB (~5 x 105 CFU/mL) and incubated for two hour s before adding the antibiotic. Selection of the RWJ-416457 and linezolid conc entration range tested was based on their
105 respective MIC values and included minimu m antimicrobial activity (0.25xMIC, 0.5xMIC, 1xMIC), efficient bacterial killing (2xMIC, 4xMIC ), as well as, maximum kill effect (8xMIC, 16xMIC) (258). In addition, a growth control (n o antibiotic) was also run simultaneously. Culture flasks were incubated a nd bacterial counts were subsequen tly determined at predefined time points for up to 24 hours. Samples (20 L) were taken directly out of the flasks, diluted in ten-fold increments and plated onto 5% sheep blood agar plates using an adopted droplet-plate method (258). Bacterial counts were determined on all countable plates (limit of bacterial quantification 15-200 CFUs) after a 20-hour incubation period. A ll constant concentration timekill curve experiments were run in triplicate on separate occasions. Changing Concentration Time-Kill Curves An in vitro syringe model was used to study th e effect of changing antibiotic concentrations on MRSA OC2878. This dynamic in vitro model was based on that used in the constant concentration experiment and consisted of eight cell cultu re flasks, each equipped with an additional syringe system. These syringe syst ems allowed simulating the human half-lives of RWJ-416457 (~24h) and linezolid (~ 5h) (95, 127) by subsequent replacement of antibioticcontaining MHB with fresh, antibio tic-free medium. To avoid dilution of the bacterial inoculum during the removal/replacement process, sterile filters (0.22 m, Millipore, Billerica, MA, USA) were placed between flasks and syringes. Prior to the start of the experiment, flasks were filled with 20mL MHB and incubated overnight to test for contamination. At the day of the experiment, flasks were sp iked with MRSA OC2878 (~5x105 CFU/mL) and incubated for 2 hours before adding the respectiv e antibiotic. Initial RWJ-416457 a nd linezolid concentrations in the changing concentration experiment were similar to those employed in the constant concentration time-kill curve and ranged from 0.25 to 16xMIC. The antibiotic inoculum was diluted during the 24-hour time course of the experiment by replacing 2.2mL RWJ-416457-
106 containing MHB every 4 hours or 4.8mL linezoli d-containing MHB every 2 hours with fresh antibiotic-free medium. Samples were taken directly out of the flasks, diluted with sterile saline solution and plated onto 5% sheep blood agar plates using an adopted droplet-plate method (258). Bacterial counts were dete rmined on all countable plates ( limit of bacteria l quantification 15-200 CFUs) after a 20-hour incubation period. All changing concentration time-kill curve experiments were run in triplicate. Drug Stability To ensure stability of RWJ-416457 and linezolid during the 24-hour time course of the experiment, samples (500 L) were taken every 8 hours directly out of the flask containing the highest antibiotic concentration (16xMI C) and immediately frozen at minus 80oC. Samples were analyzed by a validated high-performan ce liquid chromatographic-quadrupole mass spectrometric (LC-MS/MS) method. Mathematical Modeling A susceptibility-based two-compartment model (F igure 6-1) was used to characterize the constant, as well as, changing concentration time-kill curve data of both RWJ-416457 and linezolid. In this model, the overall change in the experimentally determined total number of bacteria (N) was defined as the sum of bacteria susceptible to antibiotic (NS) and insusceptible persister cells (NP) as shown in equation 6-1. PSNNN (6-1) Bacteria from both susceptibility stages can transform into each other with the transformation-rate constants ksp (h-1) and kps (h-1), respectively, as shown in Figure 6-1. The initial fraction of bacteria in the susceptible or persister stage, respectively, was defined as F (0
107 The change in number of sus ceptible bacteria over time dNS/dt could be sufficiently described by equation 6-2, PPSSdSP tdk 50 max tdg max s S SNkNkke1 CEC Ck e1 N N 1k dt dN (6-2) where kS (h-1) characterizes the growth-rate constant, kd (h-1) the natural death-rate constant, Nmax (CFU/mL) the maximum number of bacteria, dgs (h-1) the delay in the onset of growth, kmax (h-1) the maximum kill-rate constant, C (m g/mL) the antibiotic concentration, EC50 (mg/mL) the concentration necessary to produce 50% of the maximum effect and dks (h-1) the delay in the onset of kill. In comparison, the change in persister cells over time dNP/dt could be described as a function of susceptible cells enteri ng the persister stage, as well as persistent bact eria re-entering the susceptible stage and natural death of b acteria in the persister stage (equation 6-3). PdPPSSSP PNkNkNk dt dN (6-3) Data Analysis This susceptibility-based two-compartment model was simultaneously fitted to the log transformed data of the constant, as well as, ch anging concentration time -kill curve experiments using a first-order conditional estimation (F OCE) method algorithm as implemented in NONMEM 6 (Globomax, Hanover, MD, USA, ADVAN6 ). Between-experiment variability was estimated on the model parameters using expone ntial error models. The residual variability, which includes the within experimental variab ility and model mis-spec ification, was estimated using a log error model.
108 EC50 comparisons were performed using a two-sided, two-sample T-test. A P value of < 0.05 was considered statis tically significant. Model Validation Evaluation of the model performance include d diagnostic plots, Akaike information criterion (AIC), precision of th e parameter estimates, as well as, visual inspection for the quality of fit. The robustness of the final model was assessed in Wings for NONMEM 6 by a nonparametric bootstrap. In the nonparametri c bootstrap procedure, bacterial samples corresponding to one strain and concentration were sampled 1000 times with replacement from the original data set in order to obtain a new data set containing the same number of samples. The final model was fit individually to each of these new data sets and all population model parameters were estimated. Results from the successful runs were determined and median bootstrapped parameter values (including a 90% bootstrap confidence interval) were compared to the final model predicte d parameter estimates (80). Results MIC Using a modified Macrodilution broth met hod, a two-fold difference in MIC values (modes) against MRSA OC2878 was determined for RWJ-416457 (0.5 g/mL) and linezolid (1.0 g/mL), respectively. Constant Concentration Time-Kill Curves Constant concentration time-kill profiles of both RWJ-416457 and linezolid are shown in Figure 6-2A. After an initial lag phase of about 2 hours, a ~2-2.5 log reduction in bacterial counts could be observed after 24 hours of antibio tic exposure at concentr ations greater than eight times the MIC (8xMIC). Concentrations necessary to produce this maximum kill effect were approximately two-fold lower for RWJ-416457 (4 g/mL) compared to linezolid (8 g/mL).
109 Changing Concentration Time-Kill Curves Changing concentration time-kill profiles (Figure 6-2B) show that afte r an initial ~2 log kill, all linezolid concentrations fail to prevent bacterial re-growth within 8 hours of exposure. In contrast, RWJ-416457 concentrations of greater than eight times the MIC exhibit sufficient bacteriostatic activity af ter 24 hours of exposure. Drug Stability While linezolid was completely stable, a pproximately 10% of RWWJ-416457 degraded over the time course of 24 hours during the ex periment (data not shown). Hence, the degradation-rate constant of RWJ-416457 was determined (assum ing first-order degradation kinetics) and incorporated in to the mathematical model. Mathematical Modeling The final PK/PD model was capable of describi ng the constant (Figur e 6-2A), as well as, changing concentration (Figure 6-2B) time-ki ll curves of RWJ-416457 and linezolid against MRSA OC2878 reasonably well. Corresponding model parameters (MSE) are listed in Table 6-1. In the final model, Nmax, dgs and kd were estimated from the growth control data and assumed to be constant. Furthermore, kps was fixed to zero (198). Since the addition of a Hill factor did not significantly im prove the overall fit, an Emaxmodel rather than a sigmoidal Emaxmodel was used. In addition, allowing for between-experiment variability on kps, F, kmax and EC50 using a log error model did signifi cantly improve the final model fit. Model Validation Mean final model-predicted pa rameter estimates (MSE) and results of the nonparametric bootstrap runs (n=1000) are in good agreement as shown in Table 1. All parameter estimates from the final model lay within the 95% bootstra p CI. When plotted ag ainst the data, population predicted values (Fig. 3A), as well as, individu al predicted values (Fig. 3B) are uniformly and
110 randomly distributed around the line of iden tity. In addition, no trend was observed when plotting weighted residual vers us individual model predicted (F ig. 3C) and weighted residuals versus time (Fig. 3D). In combination with the results of the nonparametric bootstrap run, the diagnostic plots indicate that the mode l is robust and shows good predictability. Discussion The selection of an appropr iate dose and dosing regimen is a fundamental step for therapeutic success with any pharmacological ag ent (178). For antimicrobial agents, the selection of the best drug and dosing scheme for a specific pathogen not only increases the chances of cure while preventing toxic side effects, but also decreases the probability of the infecting pathogen becoming resi stant to the antimicrobial agent (123, 249). With a good understanding of the dose-exposur e relationship, or PK, and th e exposure-response relationship, or PD, it may be possible to identify a quantit ative link between the dose/dosing regimen on one hand, and the desired, as well as, undesired drug e ffects on the other hand. For antibiotics, this link has been established by correlating PK para meters that are based on free () plasma or serum concentrations to the MIC of the respec tive pathogen. To date, three main MIC-based PK/PD indices have been identified for antimicrobials: the cumulative pe rcentage of the dosing interval that the free drug concentration exceeds the MIC at steady-state conditions (T>MIC), the area under the free concentration-time curve at steady-state divided by the MIC (AUC/MIC) and the free peak level divided by the MIC (Cmax/MIC) (183). However, the MIC as single point in time estimate is not capable of characterizing the time course of neither growth nor antibiotic-induced kill or the antibiotic effect at concentrations besi des the MIC (188, 198, 231). In addition, the methodology to determine the actua l MIC value has not yet been internationally standardized and is a source of variability between different MIC determination methods (183). To overcome these limitations, other susceptibility breakpoints, such as, the EC50 have been
111 suggested as the PD input for PK/PD indices. The EC50 can be obtained, together with parameters characterizing bacterial growth and maximum antibiotic-induced kill, from continuous measurement of the anti biotic concentration-effect re lationship over time (time-kill curves) (11, 198, 258). In general, there are two different types of time-kill curves, based on the concentration profile used in these in vitro models, constant and changing concentration experiments (229). While constant concentratio n models represent stead y-state concentrations obtained after constant-rate infusi on, changing concentra tion models try to simulate the change in antibiotic concentrations that occurs in vivo. During changing concentration experiments, the desired concentration-time pr ofile can either be generate d manually by using syringes or automatically by employing pump systems. Ho wever, it has been shown in previous experiments, that the flow-rate necessary to simulate a RWJ-416457 half-life of approximately 24 hours is so slow that bacteria can actually grow back into th e broth only reservoir and cause contamination. In this case, the use of the pump systems is not desirable and, consequently, syringes have been used. Once the time-kill curve experiments have been performed, evaluation of the respective outcomes allows comparing the antimicrobial activity of RWJ-416457 an d linezolid against MRSA OC2878 over a wide range of concentrations. Qualitative analysis of the constant concentration time-kill curves revealed that the b acterial counts in these experiments are reduced by less than 3 log steps indicating that both RWJ-416457 and linezolid are bacteriostatic rather than bactericidal antibiotics ( 210). It could be further show n that linezolid concentrations (8g/mL) necessary to reach the maximum kill ef fect were approximately two-fold higher than the respective RWJ-416457 concentr ations (4g/mL). These tw o-fold differences in the concentration-effect relationship were consistent with the observe d two-fold differences in the
112 MICs (linezolid: 1.0g/mL, RWJ-416457: 0.5g/mL). In theory, the increased antimicrobial activity could be explained by a higher potency or a larger number of molecule s at the effect site. Yet, the molecular weights of linezolid (337.35g/mol) a nd RWJ-416457 (377.41g/mol) are not substantially different. As a result, the increased activity of RWJ-416457 against MRSA OC28768 is explained by a higher potency ra ther than the number of molecules. In order to evaluate the effect of d ecreasing concentrations (according to the physiological half-life) on the antimicrobial outcome, changing concentration time-kill curves were performed. Findings indicate that initi al RWJ-416457 concentrations of 4g/mL were sufficient to obtain bacteriostat ic activity, whereas even fou r-fold higher initi al linezolid concentrations (16g/mL) failed to prevent bacterial regrowth after 24 hours of incubation. These findings imply that, compared to linezoli d, the increased potency and prolonged half-life of RWJ-416457 may support a lower dose and/or increased dosing interv al. In order to identify an appropriate dosing regimen for RWJ-416457, tim e-kill curve-based modelling and simulation approaches can be used. Although the descrip tion of time-kill curves is mathematically somewhat more complex, PD parameters derived from this in vitro model can then be combined with in vivo PK data to simulate the antimicrobial efficacy of the respective doing regimen(s). Once a general mathematical model is established, it can be applied to the time-kill curve data of other investigational oxazolidinone s and respective outcome para meters between drugs and/or dosing regimens can be compared. Modelling and simulation approaches are, consequently, very valuable for dose selection and have been recommended by the U.S. Food and Drug Administration (FDA) as tools to streamline the drug development process (84). In fact, modelbased comparison of a new drug candidate to th e approved first-in-class representative may allow demonstrating superiority rather than non-inferiority.
113 The model that was found appropriate to simu ltaneously characterize the static, as well as, dynamic time-kill curve data of both RWJ-416457 and linezolid has structural similarities to previously described models (126, 198, 275). In th is susceptibility-based model, bacteria can exist in two different metabolic stages, an active, growing state (S) and a dormant resting stage (P) as shown in Fig. 6-1. In addition, the model allows accounting for naturally occurring deaths, saturation in growth and de lays in the onset of growth, as well as, antibiotic-induced kill. When simultaneously fitted to the data, th e final model was capable of describing the experimentally determined time-kill curve data reasonably well (Fig. 6-2). The overall model fit could be improved by incorporating drug degr adation during the 24hour course of the experiment. The final model was internally validated by nonparametric bootstrapping and showed good robustness and predictability (F ig. 6-3). Comparison of the obtained EC50 values revealed that RWJ-416457 (0.41g/mL) is approximate ly 3.4-fold more potent than the first-inclass representative linezolid (1.39g/mL). The parameter estimates obtained from this in vitro model can then be used in combination with PK data to predict clinical outcome and provide guidance for the selection of appropriate doses or dosing regimens. In conclusion, a general PK/PD model has been developed that is appropriate for characterizing the in vitro time-kill curve data of oxazolidinone antibiotics. Simultaneous fit of the developed model to static, as well as, dyna mic time-kill curve data revealed that RWJ416457 has a 3.4-fold increased in vitro potency compared to the first-in-class representative, linezolid. Combined with appropriate PK data, this model may provide valuable guidance for dose or dosing regimen selection of new, investigational oxazolidinones.
114 Table 6-1. Comparison of the final model parame ter estimates (MSE) and estimates (95%CI) from 1000 nonparametric bootstrap runs Parameter and Model Final Model Bootstrap Estimates (n=1000) Structural model ks (h-1) 1.19 (.061)a1.21 (1.11-1.35)b kmax (h-1) 1.65 (.065)a1.65 (1.51-1.82)b EC50(RWJ-416457) ( g/mL) 0.41 (.057)a0.32 (0.22-0.45)b EC50(linezolid) ( g/mL) 1.39 (.207)a1.13 (0.80-1.53)b Nmax (CFU/mL) Fixed to 3.39*109Fixed to 3.39*109 dgs (h-1) Fixed to 0.24 Fixed to 0.24 dks (h-1) 0.50 (.049)a0.54 (0.47-0.63)b ksp (h-1) 0.004 (.002)a0.001 (0.000-0.003)b kps (h-1) Fixed to zero Fixed to zero kd (h-1) Fixed to 0.015 Fixed to 0.015 F 0.83 (.023)a0.83 (0.79-0.88)b Variance model (ks) 0.013 (.005)a0.11 (Nmax) 0.219 (.082)a0.51 (dgs) 0.184 (.071)a0.43 (dks) 0.079 (.029)a0.26 Residual variability 0.29 (.027)a0.53 aMSE. b95%CI
115 Figure 6-1. Susceptibility-based two-subpopulation model. S = pool of metabolically active and self-replicating bacteria. P = pool of dormant persister cells. kS = growth-rate constant. kd = natural death-rate constant. ksp = transfer rate-constant from active to resting stage. kps = transfer rate-constant from resting to active stage.
116 Figure 6-2. Simultaneous curve fits of the susc eptibility-based two-compartment model to the experimental data. A) constant concen tration time-kill curve of RWJ-416457 against MRSA, B) changing concentration (t1/2~24h) of RWJ-416457 against MRSA, C) constant concentration time-kill curve of linezolid against MRSA, and D) changing concentration (t1/2~5h) of linezolid against MRSA at initial concentrations of 0.25xMIC ( ), 0.5xMIC ( ), 1xMIC ( ), 2xMIC ( ), 4xMIC ( ), 8xMIC ( ) and 16xMIC ( ) plus a growth control ( ). Symbols represent the experimental data, lines the model predicted curve fits.
117 Figure 6-3. Basic diagnostic pl ots. A) observed versus popula tion model predicted, B) observed versus individual model predicted, C) wei ghted residual versus individual model predicted and D) weighted residuals versus time. The dotted line represents the line of identity; the solid line represents the linear regression line.
118 CHAPTER 7 DISCUSSION Since the discovery of the penicillin by Si r Alexander Flem ing in 1928, antibiotics have effectively been used for decades to treat bact erial infections (208, 274). However, due to an over-/misuse of first choice antibiotics (e.g. for minor infections or in food-producing animals), the failure to complete treatment, and the substantial increase in the availability and ease of travel within/b etween countries has contributed to a rapid emergence and spread of resistance during the second half of the 20th century (165, 171, 270). Resistance against antibiotics, as stated in the annual CDC reports and results of the SENTRY study, is a major cause of morbidity and mortality in the Un ited States and all over the world (70, 106, 118, 163, 257, 271) and accounts for excess days of illness and hospitalization (257), as well as, increased health care costs (209). In order to increase the chances of clinical success and mini mize the risk of resistance development, it is critically important to impr ove antibiotic therapy by optimizing current dosing regimens and developing newer and more potent antibiotics. The most comprehensive approach of achieving this goal is to apply PK/PD pr inciples throughout the optimization/development process. On the other hand, disregard of these principles may result in conflicting outcomes, failure in later and more costly clinical stages and potential harm to the patient. Chances of success are increased when these principles are employed early on in the drug development process. For example, commercially available pr otein supplements are used in pre-clinical test systems to produce and modify protein binding. However, if the actual binding in these in vitro test systems represents phys iological binding conditions is frequently not determined. The first specific goal of this thesis work was to evaluate this approach by experimentally measuring free, unbound ceftriaxone (protein bind ing: 83-96%) (244, 24 5, 277) and ertapenem
119 (protein binding: 84-96%) (34) concentrations using in vitro microdialysis. Findings of this study clearly indicate that the protein binding values achieved by supplementing with commercially available albumins can substantially differ from those in serum/plasma. It was concluded that in order to avoid experimental misconduct and further misinterpretation of the effect of protein binding on the antimicrobial activity, free, unbound concentrations need to be measured in the actual in vitro test system. Traditionally, the effect of protein binding is taken into account by linking free, unbound concentrations in plasma to the respective PD outcome. However, since most bacterial infections are located in the ex tracellular fluid of tissues, a mo re comprehensive approach would be to measure free, active concentrations in the interstitial space fluid of tissues rather than the blood. Microdialysis is currently the only technique that can pr ovide this information and has been extensively been used in tissue distribution studies. The second specific goal of this thesis work wa s to measure free, activ e concentrations in the interstitial space fluid of muscle and subcutaneous adipose tissue of six healthy subjects following a single intravenous infusion of 1g ertapenem using clinical microdialysis. Results of this microdialysis study indicate that free, unbound ertapenem concentrations in th e interstitial space fluid of muscle are comparable to free concentrations in plasma but are higher than those in subcutaneous adipose tissue. Nevertheless, free concentrations in both tissues were sufficient to exceed the MICs of the most prevalent skin and skin structure pathogens. These findings support the clinical use of the approved dosing regimen of 1g ertapenem BID for the treatment of skin and skin structure infections. To date, PK information, determined in e.g. mi crodialysis experiments, is frequently linked to the MIC in order to evaluate dose-concentrat ion-response relationships. However, as stated
120 previously, the MIC suffers from numerous draw backs. To overcome these limitations, other approaches, such as, time-kill curves have been suggested for evaluating the PD outcome. In order to characterize the experimental time-kill curves, an appropriate PK/PD model can be simultaneously fitted to the data. Once determ ined, the PD parameters can be linked to PK profiles of different doses and dosing regimens to predict and compare the respective clinical outcome. The third specific goal of this thesis work was to establish and validate an appropriate PK/PD model that allows charac terizing the antimicrobial activity of oxazolidinones, determined in time-kill curve experiments, against MRSA. Re sults of this study indicat e that the identified susceptibility-based two-subpopulation model was appropriate to describe the oxazolidinone effect on MRSA over time. When simultaneously fitted to the data, comparison of the corresponding EC50 values revealed that RWJ-416457 is approximately 3.4-fold more potent than the first-in-class representative linezolid. In combination with appropriate PK data, this model can be used to guide the pre-clinic al development proce ss of new oxazolidinone antibiotics and, subsequently, to select an appropriat e dose for clinical trials. In summary, the proper application of PK/P D principles throughout drug development and in the clinic supports the estab lishment of effective dosing regimens and helps to prevent the emergence of resistance. However, the comple x dynamic interaction between a host, a pathogen and the infection process has not been comple tely understood (236). To date, the immune system is seen predominantly as a defence mech anism of the body against in fections. Little is known about its impact on the manifestation of clinical symptoms. Therefore, in future approaches patient-, disease a nd immune status should be take n into greater consideration (Figure 6-5) (166, 167).
121 Most comprehensive assessments of interact ions between pathogen, host and antibiotic can be achieved by integrating i ndividual PK/PD profiles into modeling and support the establishment of effective dosage regimens fo r treatment (73, 174, 196). However, clinical practice often differs from this ideal situation when treatment is started out empirically to improve the status of the patient (131, 233). Ye t, for success in treatment, eradication of the pathogen should be regarded as the primary endpoint in order to avoid emergence and spread of resistant strains (58, 93). Combined efforts will be needed to achieve this goal and meet the challenge of multi-drug-resistant microbe threats in the clinic as well as in the community. Hence, not only rethinking of antimicrobial tr eatment and prescribing standards but also a change in patient care/management will be nece ssary (60, 86, 265). In drug development, PK and PD criteria can be used to predict bact eriologic efficacy. However, they should be confirmed during all phases of antimicrobial development and throughout clinical use in response to changing resistan ce patterns (7, 57, 113, 131, 238). Identifying new compounds is one potent strategy but not th e only one. If new compounds are derived from existing drug classes, cross-resi stance can occur (7, 267). Hence, it is essential that existing drug classes are used in optimiz ed ways. For example, new dosage forms of existing compounds (e.g. liposomal antimicrobial agents), drug combinations and advanced modeling and simulation tools need to be develo ped in order the improve PK/PD and establish effective dosing regimens (86, 181, 220).
122 Figure 7-1. Interplay between pharmacokine tics, pharmacodynamics, patient and disease
123 LIST OF REFERENCES 1. Allegranzi, B., A. Cazzadori, G. Di Perri, S. Bonora, M. Berti, L. Franchino, A. Biglino, A. Cipriani, and E. Concia. 2000. Concentratio ns of single-dose meropenem (1 g iv) in bronchoalveolar lavage and ep ithelial lining fluid. J Antimicrob Chemother 46:319-22. 2. Ambrose, P. G. 2006. Monte Carlo simulati on in the evaluation of susceptibility breakpoints: predicting the futu re: insights from the society of infectious diseases pharmacists. Pharmacotherapy 26:129-34. 3. Ambrose, P. G., S. M. Bhavnani, B. B. Ci rincione, M. Piedmonte, and T. H. Grasela. 2003. Gatifloxacin and the elderly: pharm acokinetic-pharmacodynamic rationale for a potential age-related dose reducti on. J Antimicrob Chemother 52:435-40. 4. Ambrose, P. G., S. M. Bhavnani, and R. C. Owens, Jr. 2003. Clinical pharmacodynamics of quinolones. Infect Di s Clin North Am 17:529-43. 5. Ambrose, P. G., and D. M. Grasela. 2000. Th e use of Monte Carlo simulation to examine pharmacodynamic variance of drugs: fluoroquinolone pharmacodynamics against Streptococcus pneumoniae. Dia gn Microbiol Inf ect Dis 38:151-7. 6. Andes, D., posting date. Animal Model PK/P D: A Tool for Drug Development. [Online.] 7. Andes, D. 2001. Pharmacokinetic and pharm acodynamic properties of antimicrobials in the therapy of respiratory tract infe ctions. Curr Opin Infect Dis 14:165-72. 8. Andes, D., and W. A. Craig. 2002. Phar macodynamics of the new fluoroquinolone gatifloxacin in murine thigh and lung infe ction models. Antimicrob Agents Chemother 46:1665-70. 9. Bahlmann, L., M. Misfeld, S. Klaus, A. Lep tien, M. Heringlake, P. Schmucker, H. H. Sievers, U. Ungerstedt, and E. G. Kraatz. 2004. Myocardial redox state during coronary artery bypass grafting assessed with micr odialysis. Intensive Care Med 30:889-94. 10. Balaban, N. Q., J. Merrin, R. Chait, L. Kowalik, and S. Leibler. 2004. Bacterial persistence as a phenotypic switch. Science 305:1622-5. 11. Bantar, C., C. Schell, G. Posse, A. Li mansky, V. Ballerini, and L. Mobilia. 2008. Comparative time-kill study of doxycycline, tigecycline, sulbactam, and imipenem against several clones of Acinetobacter baum annii. Diagn Microbiol Infect Dis. 61:30914 12. Barre, J., J. M. Chamouard, G. Houin, and J. P. Tillement. 1985. Equilibrium dialysis, ultrafiltration, and ultracentrifugation compared for determining the plasma-proteinbinding characteristics of valpro ic acid. Clin Chem 31:60-4.
124 13. Barre, J., F. Didey, F. Delion, and J. P. Tillement. 1988. Problems in therapeutic drug monitoring: free drug level monito ring. Ther Drug Monit 10:133-43. 14. Benfeldt, E., S. H. Hansen, A. Volund, T. Menne, and V. P. Shah. 2007. Bioequivalence of topical formulations in humans: evaluation by dermal microdialysis sampling and the dermatopharmacokinetic method. J Invest Dermatol 127:170-8. 15. Benfeldt, E., J. Serup, and T. Menne. 1999. Effect of barrier pert urbation on cutaneous salicylic acid penetration in human skin: in vivo pharmacokinetics using microdialysis and non-invasive quantification of ba rrier function. Br J Dermatol 140:739-48. 16. Bito, L., H. Davson, E. Levin, M. Murray, and N. Snider. 1966. The concentrations of free amino acids and other electrolytes in cere brospinal fluid, in vivo dialysate of brain, and blood plasma of the dog. J Neurochem 13:1057-67. 17. Boelsma, E., C. Anderson, A. M. Karlsson, and M. Ponec. 2000. Microdialysis technique as a method to study the percutaneous penetr ation of methyl nico tinate through excised human skin, reconstructed epidermis, and human skin in vivo. Pharm Res 17:141-7. 18. Bolinder, J., E. Hagstrom, U. Ungerste dt, and P. Arner. 1989 Microdialysis of subcutaneous adipose tissue in vivo for continuous glucose monitoring in man. Scand J Clin Lab Invest 49:465-74. 19. Borg, N., E. Gotharson, E. Benfeldt, L. Groth, and L. Stahle. 1999. Distribution to the skin of penciclovir after oral famcicl ovir administration in healthy volunteers: comparison of the suction blister technique and cutaneous microdialysis. Acta Derm Venereol 79:274-7. 20. Boswell, F. J., J. P. Ashby, J. M. Andrews, and R. Wise. 2002. Effect of protein binding on the in vitro activity and pharmacodynamics of faropenem. J Antimicrob Chemother 50:525-32. 21. Boutsiouki, P., J. P. Thompson, and G. F. Clough. 2001. Effects of local blood flow on the percutaneous absorption of the organophosphorus compound malathion: a microdialysis study in man. Arch Toxicol 75:321-8. 22. Box, G. E. P. 1979. Robustness in the strate gy of scientific model building. Academic Press, New York. 23. Bratzler, D. W., and P. M. Houck. 2004. Antimicrobial prophylaxis for surgery: an advisory statement from the National Surgical Infection Prevention Project. Clin Infect Dis 38:1706-15. 24. Bratzler, D. W., and P. M. Houck. 2005. Antimicrobial prophylaxis for surgery: an advisory statement from the National Surgical Infection Prevention Project. Am J Surg 189:395-404.
125 25. Brunner, M., P. Dehghanyar, B. Seigfried, W. Martin, G. Menke, and M. Muller. 2005. Favourable dermal penetration of diclofenac af ter administration to the skin using a novel spray gel formulation. Br J Clin Pharmacol 60:573-7. 26. Brunner, M., H. Derendorf, and M. Mu ller. 2005. Microdialysis for in vivo pharmacokinetic/pharmacodynamic characterizati on of anti-infective drugs. Curr Opin Pharmacol 5:495-9. 27. Brunner, M., U. Hollenstein, S. Delacher, D. Jager, R. Schmid, E. Lackner, A. Georgopoulos, H. G. Eichler, and M. Muller. 1999. Distribution and antimicrobial activity of ciprofloxacin in human soft tissues. Antimicrob Agents Chemother 43:1307-9. 28. Brunner, M., and O. Langer. 2006. Microdialysi s versus other techniques for the clinical assessment of in vivo tissue dr ug distribution. Aaps J 8:E263-71. 29. Brunner, M., H. Stabeta, J. G. Moller, C. Schrolnberger, B. Erovic, U. Hollenstein, M. Zeitlinger, H. G. Eichler, and M. Muller. 2002. Target site concentrations of ciprofloxacin after single intravenous and oral doses. Antimicrob Agents Chemother 46:3724-30. 30. Bur, A., C. Joukhadar, N. Klein, Herkner, G. Mitulovic, R. Schmid, E. Agneter, M. Muller, and M. Brunner. 2005. Effect of exer cise on transdermal nicotine release in healthy habitual smokers. Int J Clin Pharmacol Ther 43:239-43. 31. Burian, M., I. Tegeder, M. Seegel, and G. Geisslinger. 2003. Peripheral and central antihyperalgesic effects of di clofenac in a model of human inflammatory pain. Clin Pharmacol Ther 74:113-20. 32. Burke, J. P. 2003. Infection control a probl em for patient safet y. N Engl J Med 348:6516. 33. Burkhardt, O., K. Borner, N. von der Hoh, P. Koppe, M. W. Pletz, C. E. Nord, and H. Lode. 2002. Singleand multiple-dose pharmacokinetics of linezolid and co-amoxiclav in healthy human volunteers. J Antimicrob Chemother 50:707-12. 34. Burkhardt, O., M. Brunner, S. Schmidt, M. Grant, Y. Tang, and H. Derendorf. 2006. Penetration of ertapenem into skeletal muscle and subcutaneous adi pose tissue in healthy volunteers measured by in vivo microdial ysis. J Antimicrob Chemother. 58:632-6 35. Burkhardt, O., J. Majcher-Peszynska, K. Borner, R. Mundkowski, B. Drewelow, H. Derendorf, and T. Welte. 2005. Penetration of ertapenem into different pulmonary compartments of patients undergoing l ung surgery. J Clin Pharmacol 45:659-65.
126 36. Cafini, F., L. Aguilar, N. Gonzalez, M. J. Gimenez, M. Torrico, L. Alou, D. Sevillano, P. Vallejo, and J. Prieto. 2007. In vitro effect of the presence of human albumin or human serum on the bactericidal activity of daptomycin against strains with the main resistance phenotypes in Gram-positives. J Antimicrob Chemother 59:1185-9. 37. Campion, J. J., P. J. McNamara, and M. E. Evans. 2005. Pharmacodynamic modeling of ciprofloxacin resistance in Staphylococcus aureus. Antimicrob Agents Chemother 49:209-19. 38. Cawello, W. 1997. Connection of pharmacokinetics and pharmacodynamics--how does it work? Int J Clin Phar macol Ther 35:414-7. 39. Cazzola, M., M. G. Matera, and P. Noschese 2000. Parenteral antibiotic therapy in the treatment of lower respiratory tract infections. Strategies to minimize the development of antibiotic resistance. Pulm Pharmacol Ther 13:249-56. 40. Cha, R., W. J. Brown, and M. J. Rybak. 2003. Bactericidal activi ties of daptomycin, quinupristin-dalfopristin and linezolid against vancomycin-resistant Staphylococcus aureus in an in vitro pharmacodynamic mode l with simulated endocardial vegetations. Antimicrob Agents Chemother 47:3960-3. 41. Cha, R., and M. J. Rybak. 2004. Influence of protein binding under c ontrolled conditions on the bactericidal activity of daptomycin in an in vitro pharmacodynamic model. J Antimicrob Chemother 54:259-62. 42. Chang, S., D. M. Sievert, J. C. Hageman, M. L. Boulton, F. C. Tenover, F. P. Downes, S. Shah, J. T. Rudrik, G. R. Pupp, W. J. Brown, D. Cardo, and S. K. Fridkin. 2003. Infection with vancomycin-resistant Sta phylococcus aureus containing the vanA resistance gene. N Engl J Med 348:1342-7. 43. Chao, C. C., H. Y. Sun, Y. C. Chang, and S. T. Hsieh. 2008. Painful neuropathy with skin denervation after prolonged us e of linezolid. J Neurol Ne urosurg Psychiatry 79:97-9. 44. Chaurasia, C. S., M. Muller, E. D. Bashaw, E. Benfeldt, J. Bolinder, R. Bullock, P. M. Bungay, E. C. DeLange, H. Derendorf, W. F. Elmquist, M. Hammarlund-Udenaes, C. Joukhadar, D. L. Kellogg, Jr., C. E. Lunte, C. H. Nordstrom, H. Rollema, R. J. Sawchuk, B. W. Cheung, V. P. Shah, L. Stahle, U. Ungerstedt, D. F. Welty, and H. Yeo. 2007. AAPS-FDA Workshop White Paper: microdialysis principles, application, and regulatory perspectives. J Clin Pharmacol 47:589-603. 45. Christensen, H., M. Baker, G. T. Tucker and A. Rostami-Hodjeg an. 2006. Prediction of plasma protein binding displa cement and its implications fo r quantitative assessment of metabolic drug-drug interactions from in vitro data. J Pharm Sci 95:2778-87. 46. Chulavatnatol, S., B. Chindavijak, N. Chierakul, and D. Klomsawat. 2003. Pharmacokinetics of ofloxacin in drug-resistant tuberculosis. J Me d Assoc Thai 86:781-8.
127 47. Colburn, W. A. 2003. Biomarkers in drug discovery and development: from target identification through drug marke ting. J Clin Pharmacol 43:329-41. 48. Collison, M. E., P. J. Stout, T. S. Glushko, K. N. Pokela, D. J. Mullins-Hirte, J. R. Racchini, M. A. Walter, S. P. Mecca, J. Rundquist, J. J. Allen, M. E. Hilgers, and T. B. Hoegh. 1999. Analytical characterization of el ectrochemical biosensor test strips for measurement of glucose in low-volume inters titial fluid samples. Clin Chem 45:1665-73. 49. Craig, W. A. 2001. Does the dose matter? Clin Infect Dis 33:233-7. 50. Craig, W. A. 2001. The hidden impact of an tibacterial resistance in respiratory tract infection. Re-evaluating current antibio tic therapy. Respir Med 95:12-9; 26-7. 51. Craig, W. A. 1995. Interrelationship betw een pharmacokinetics and pharmacodynamics in determining dosage regimens for broad-spectrum cephalosporins. Diagn Microbiol Infect Dis 22:89-96. 52. Craig, W. A. 2003, posting date. Pharmac odynamics of Antimicrobials in Animal Models. [Online.] 53. Craig, W. A. 2002. Pharmacodynamics of antimicrobials: general concepts and applications, p. 1-22, Antimicrobial Pharmacodynamics in Theory and Clinical Practice. Marcel-Dekker, New York. 54. Critchley, I. A., D. F. Sahm, C. Thornsberry, R. S. Blosser-Middleton, M. E. Jones, and J. A. Karlowsky. 2002. Antimicrobial susceptibili ties of Streptococcus pyogenes isolated from respiratory and skin and soft tissue infections: United States LIBRA surveillance data from 1999. Diagn Micr obiol Infect Dis 42:129-35. 55. Cross, S. E., C. Anderson, and M. S. Roberts. 1998. Topical penetr ation of commercial salicylate esters and salts using human isolated skin and clin ical microdialysis studies. Br J Clin Pharmacol 46:29-35. 56. Cunha, B. A. 2002. Ertapenem. A review of its microbiologic, pharmacokinetic and clinical aspects. Drugs Today (Barc) 38:195-213. 57. Dagan, R. 2003. Achieving b acterial eradication using pharmacokinetic/pharmacodynamic principles. Int J Infect Dis 7:21-6. 58. Dagan, R., K. P. Klugman, W. A. Craig, and F. Baquero. 2001. Evidence to support the rationale that bacterial eradication in respir atory tract infection is an important aim of antimicrobial therapy. J Antimicrob Chemother 47:129-40. 59. Dasgupta, A. 2007. Usefulness of monitoring free (unbound) concentrations of therapeutic drugs in pa tient management. Clin Chim Acta 377:1-13.
128 60. Davenport, L. A., P. G. Davey, and J. S. Ker. 2005. An outcome-based approach for teaching prudent antimicrobial prescribing to undergraduate medical students: report of a Working Party of the British Society for Antimicrobial Chemotherapy. J Antimicrob Chemother 56:196-203. 61. Day, R. M., M. Harbord, A. Forbes, and A. W. Segal. 2001. Cantharidin blisters: a technique for investigating leukocyte traffick ing and cytokine production at sites of inflammation in humans. J Immunol Methods 257:213-20. 62. de Hoog, M., J. W. Mouton, and J. N. va n den Anker. 2005. New dosing strategies for antibacterial agents in the neonate Semin Fetal Neonatal Med 10:185-94. 63. de Lange, E. C., A. G. de Boer, and D. D. Breimer. 2000. Methodological issues in microdialysis sampling for pharmacokineti c studies. Adv Drug Deliv Rev 45:125-48. 64. Dehghanyar, P., C. Burger, M. Zeitlinger, F. Is linger, F. Kovar, M. Muller, C. Kloft, and C. Joukhadar. 2005. Penetration of linezolid into soft tissues of hea lthy volunteers after single and multiple doses. Antimicrob Agents Chemother 49:2367-71. 65. Del Mar Fernandez de Gatta Garcia, M., N. Revilla, M. V. Calvo, A. Dominguez-Gil, and A. Sanchez Navarro. 2006. Pharmacokinetic/ pharmacodynamic analysis of vancomycin in ICU patients. Intens ive Care Med. 33:279-85 66. Delgado, J. M., F. V. DeFeudis, R. H. Roth, D. K. Ryugo, and B. M. Mitruka. 1972. Dialytrode for long term intracerebral perfusion in awake monkeys. Arch Int Pharmacodyn Ther 198:9-21. 67. Derendorf, H., Gramatt, T., Schfer, H. G. 2002. Pharmakodynamische Modelle, p. 278282, Pharmakokinetik, Volume 2 ed. Wissens chaftliche Verlagsgesellschaft mbH, Stuttgart. 68. Derendorf, H., L. J. Lesko, P. Chaikin, W. A. Colburn, P. Lee, R. Miller, R. Powell, G. Rhodes, D. Stanski, and J. Venitz. 2000. Pharmacokinetic/pharmacodynamic modeling in drug research and development. J Clin Pharmacol 40:1399-418. 69. Derendorf, H., and B. Meibohm. 1999. M odeling of pharmacokinetic/pharmacodynamic (PK/PD) relationships: concepts a nd perspectives. Pharm Res 16:176-85. 70. Diekema, D. J., M. A. Pfaller, F. J. Schmitz, J. Smayevsky, J. Bell, R. N. Jones, and M. Beach. 2001. Survey of infections due to Staphylococcus species: frequency of occurrence and antimicrobial susceptibility of isolates collected in the United States, Canada, Latin America, Europe, and the Western Pacific region for the SENTRY Antimicrobial Surveillance Program, 1997-1999. Clin Infect Dis 32 Suppl 2:S114-32.
129 71. DiNubile, M. J., and B. A. Lipsky. 2004. Co mplicated infections of skin and skin structures: when the infection is more than skin deep. J Antimicrob Chemother 53 Suppl 2:ii37-50. 72. Dresser, L. D., and M. J. Rybak. 1998. The pharmacologic and bacteriologic properties of oxazolidinones, a new class of synthetic antimicrobials. Pharmacotherapy 18:456-62. 73. Drusano, G. 2001. Pharmacodynamic and pharmacokinetic considerations in antimicrobial selection: focus on telithr omycin. Clin Microb iol Infect 7:24-9. 74. Drusano, G. 2001. Pharmacodynamic and pharmacokinetic considerations in antimicrobial selection: focus on telithromyc in. Clin Microbiol In fect 7 Suppl 3:24-9. 75. Drusano, G. L. 2004. Antimicrobial pharmacodyna mics: critical intera ctions of 'bug and drug'. Nat Rev Microbiol 2:289-300. 76. Drusano, G. L., D. E. Johnson, M. Rosen, and H. C. Standiford. 1993. Pharmacodynamics of a fluoroquinolone antimic robial agent in a neutropenic rat model of Pseudomonas sepsis. Antimic rob Agents Chemother 37:483-90. 77. Drusano, G. L., S. L. Preston, C. Hardalo, R. Hare, C. Banfield, D. Andes, O. Vesga, and W. A. Craig. 2001. Use of preclinical data for selection of a phase II/III dose for evernimicin and identification of a prec linical MIC breakpoint. Antimicrob Agents Chemother 45:13-22. 78. Ederoth, P., K. Tunblad, R. Bouw, C. J. Lundberg, U. Ungerstedt, C. H. Nordstrom, and M. Hammarlund-Udenaes. 2004. Blood-brain barrier transport of morphine in patients with severe brain trauma. Br J Clin Pharmacol 57:427-35. 79. Engstrom, M., A. Polito, P. Reinstrup, B. Ro mner, E. Ryding, U. Ungerstedt, and C. H. Nordstrom. 2005. Intracerebral microdialysis in severe brain trauma: the importance of catheter location. J Neurosurg 102:460-9. 80. Ette, E. I. 1997. Stability a nd performance of a population pharmacokinetic model. J Clin Pharmacol 37:486-95. 81. FDA May 1, 2007, posting date. Bioequivale nce of Topical Dermatological Products. [Online.] 82. FDA February 21, 2003, posting date. CD ER News Along the Pike. [Online.] 83. FDA March 8, 2001, posting date. Guidance for Industry: Bioavailability and Bioequivalence Studies for Orally Ad ministered Drug Products General Considerations. [Online.]
130 84. FDA March 16 2004, posting date. Innovation or Stagnation Challenge and Opportunity on the Critical Path to Ne w Medical Products. [Online.] 85. FDA July 8, 2003, posting date. Summary of an FDA workshop, "Innovative Systems for Delivery of Drugs and Biologics: Scientific Clinical and Regulatory Challenges". [Online.] 86. Finch, R. G. 2004. Introduction: standards of antibacterial performance. Clin Microbiol Infect 10:1-5. 87. Foleno, B. D., D. Abbanat, R. M. Goldschmidt, R. K. Flamm, S. D. Paget, G. C. Webb, E. Wira, M. J. Macielag, and K. Bush. 2007. In vitro antibacterial activity of the pyrrolopyrazolyl-substituted oxazolidinone RWJ-416457. Antimicrob Agents Chemother 51:361-5. 88. Fontana, R., M. Aldegheri, M. Ligozzi, G. Lo Cascio, and G. Cornaglia. 1998. Interaction of ceftriaxone with penicillin-binding proteins of Esch erichia coli in the presence of human serum albumin. J Antim icrob Chemother 42:95-8. 89. Forrest, A., D. E. Nix, C. H. Ballow, T. F. Goss, M. C. Birmingham, and J. J. Schentag. 1993. Pharmacodynamics of intravenous cipr ofloxacin in serious ly ill patients. Antimicrob Agents Chemother 37:1073-81. 90. Frimodt-Moller, N. 2002. Correlation be tween pharmacokinetic/pharmacodynamic parameters and efficacy for antibiotics in th e treatment of urinary tract infection. Int J Antimicrob Agents 19:546-53. 91. Fuchs, P. C., A. L. Barry, and S. D. Br own. 2001. In vitro activiti es of ertapenem (MK0826) against clinical bacter ial isolates from 11 North American medical centers. Antimicrob Agents Chemother 45:1915-8. 92. Fung, H. B., J. Y. Chang, and S. Kuczynski. 2003. A practical guide to the treatment of complicated skin and soft tis sue infections. Drugs 63:1459-80. 93. Garau, J. 2003. Why do we need to eradicate pa thogens in respiratory tract infections? Int J Infect Dis 7:5-12. 94. Gattringer, R., E. Urbauer, F. Traunmuller, M. Zeitlinger, P. Dehghanyar, P. Zeleny, W. Graninger, M. Muller, and C. Joukhadar. 2004. Pharmacokinetics of telithromycin in plasma and soft tissues after single-dose ad ministration to healthy volunteers. Antimicrob Agents Chemother 48:4650-3. 95. Gee, T., R. Ellis, G. Marshall, J. Andrews, J. Ashby, and R. Wise. 2001. Pharmacokinetics and tissue penetration of linezolid following multiple oral doses. Antimicrob Agents Chemother 45:1843-6.
131 96. Giurini, J. M., and T. E. Lyons. 2005. Di abetic foot complications: diagnosis and management. Int J Low Extrem Wounds 4:171-82. 97. Goldstein, E. J., D. M. Citron, S. Hunt Ge rardo, M. Hudspeth, and C. V. Merriam. 1998. Activities of HMR 3004 (RU 64004) and HMR 3647 (RU 66647) compared to those of erythromycin, azithromycin, clarithromycin, r oxithromycin, and eight other antimicrobial agents against unusual aerobic and anaerobic hu man and animal bite pathogens isolated from skin and soft tissue infections in humans. Antimicrob Agents Chemother 42:112732. 98. Gordien, J. B., E. Boselli, C. Fleureau, B. Allaouchiche, G. Janvier, O. Lalaude, M. C. Saux, and D. Breilh. 2006. Determination of free ertapenem in plasma and bronchoalveolar lavage by high-performance liquid chromatography with ultraviolet detection. J Chromatogr B Analyt Technol Biom ed Life Sci 830:218-23. 99. Graham, D. R., C. Lucasti, O. Malafaia, R. L. Nichols, P. Holtom, N. Q. Perez, A. McAdams, G. L. Woods, T. P. Ceesay, a nd R. Gesser. 2002. Ertapenem once daily versus piperacillin-tazobactam 4 times per day for treatment of complicated skin and skinstructure infections in ad ults: results of a prospec tive, randomized, double-blind multicenter study. Clin Infect Dis 34:1460-8. 100. Gumbo, T., A. Louie, M. R. Deziel, L. M. Parsons, M. Salfinger, and G. L. Drusano. 2004. Selection of a moxifloxacin dose that s uppresses drug resistance in Mycobacterium tuberculosis, by use of an in vitro pharm acodynamic infection model and mathematical modeling. J Infect Dis 190:1642-51. 101. Heinze, A., and U. Holzgrabe. 2006. Determin ation of the extent of protein binding of antibiotics by means of an automated c ontinuous ultrafiltration method. Int J Pharm 311:108-12. 102. Herkner, H., M. R. Muller, N. Kreischitz, B. X. Mayer, M. Frossard, C. Joukhadar, N. Klein, E. Lackner, and M. Muller. 2002. Closed -chest microdialysis to measure antibiotic penetration into human lung tissue. Am J Respir Crit Care Med 165:273-6. 103. Herrera, A. M., D. O. Scott, and C. E. Lunte. 1990. Microdialysis sampling for determination of plasma protein bi nding of drugs. Pharm Res 7:1077-81. 104. Herrero, I. A., N. C. Issa, and R. Patel. 2002. Nosocomial spread of linezolid-resistant, vancomycin-resistant Enterococcus faecium. N Engl J Med 346:867-9. 105. Heyneman, C. A., C. Lawless-Liday, and G. C. Wall. 2000. Oral versus topical NSAIDs in rheumatic diseases: a comparison. Drugs 60:555-74.
132 106. Hoban, D. J., G. V. Doern, A. C. Fluit, M. Roussel-Delvallez, and R. N. Jones. 2001. Worldwide prevalence of antimicrobial re sistance in Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis in the SENTRY Antimicrobial Surveillance Program, 1997-1999. Clin Infect Dis 32 Suppl 2:S81-93. 107. Hocht, C., J. A. Opezzo, and C. A. Taira. 2004. Microdialysis in drug discovery. Curr Drug Discov Technol 1:269-85. 108. Hodges, G. J., K. Zhao, W. A. Kosiba, and J. M. Johnson. 2006. The involvement of nitric oxide in the cutaneous vasoconstrictor response to local cooling in humans. J Physiol 574:849-57. 109. Holford, N. H., and L. B. Sheiner. 1982. Kinetics of pharmacologic response. Pharmacol Ther 16:143-66. 110. Holford, N. H., and L. B. Sheiner. 1981. Pharmacokinetic and pharmacodynamic modeling in vivo. Crit Rev Bioeng 5:273-322. 111. Huang, Y., and Z. Zhang. 2004. Binding study of drug with bovine serum album using a combined technique of microdialysis with flow-injection chemiluminescent detection. J Pharm Biomed Anal 35:1293-9. 112. Hutschala, D., K. Skhirtladze, C. Kinstner, B. Mayer-Helm, M. Muller, E. Wolner, and E. M. Tschernko. 2007. In vivo microdialysis to m easure antibiotic penetration into soft tissue during cardiac surgery. Ann Thorac Surg 84:1605-10. 113. Jacobs, M. R. 2003. How can we predict bacter ial eradication? Int J Infect Dis 7:13-20. 114. Jain, R. K. 1994. Barriers to drug delivery in solid tumors. Sci Am 271:58-65. 115. Jain, R. K. 1987. Transport of molecules in the tumor interstitium: a review. Cancer Res 47:3039-51. 116. Jansson, K., M. Jansson, M. Andersson, A. Magnuson, U. Ungerstedt, and L. Norgren. 2005. Normal values and differences betw een intraperitoneal and subcutaneous microdialysis in patients after non-complicated gastrointestinal surg ery. Scand J Clin Lab Invest 65:273-81. 117. Jansson, P. A., J. Fowelin, U. Smit h, and P. Lonnroth. 1988. Characterization by microdialysis of intracellular glucose level in subcutaneous tissue in humans. Am J Physiol 255:E218-20. 118. Jones, R. N., J. T. Kirby, M. L. Beach, D. J. Biedenbach, and M. A. Pfaller. 2002. Geographic variations in activ ity of broad-spectrum beta-lactams against Pseudomonas aeruginosa: summary of the worldwide SENT RY Antimicrobial Surveillance Program (1997-2000). Diagn Microbiol Infect Dis 43:239-43.
133 119. Joukhadar, C., P. Dehghanyar, F. Traunm uller, R. Sauermann, B. Mayer-Helm, A. Georgopoulos, and M. Muller. 2005. Increase of microcirculatory blood flow enhances penetration of ciprofloxacin into soft tissue. Antimicrob Agents Chemother 49:4149-53. 120. Joukhadar, C., H. Derendorf, and M. Mulle r. 2001. Microdialysis. A novel tool for clinical studies of anti -infective agents. Eur J Clin Pharmacol 57:211-9. 121. Joukhadar, C., N. Klein, R. M. Mader, C. Sc hrolnberger, B. Rizovs ki, E. Heere-Ress, H. Pehamberger, N. Strauchmann, B. Janse n, and M. Muller. 2001. Penetration of dacarbazine and its active me tabolite 5-aminoimidazole-4-carboxamide into cutaneous metastases of human maligna nt melanoma. Cancer 92:2190-6. 122. Joukhadar, C., and M. Muller. 2005. Microdialysis: current applications in clinical pharmacokinetic studies and its potential role in the future. Clin Pharmacokinet 44:895913. 123. Jumbe, N., A. Louie, R. Leary, W. Liu, M. R. Deziel, V. H. Tam, R. Bachhawat, C. Freeman, J. B. Kahn, K. Bush, M. N. Dudley, M. H. Miller, and G. L. Drusano. 2003. Application of a mathematical model to prevent in vivo amplification of antibioticresistant bacterial popula tions during therapy. J Clin Invest 112:275-85. 124. Kaplan, A. H., P. H. Gilligan, and R. R. Facklam. 1988. Recovery of resistant enterococci during vancomycin prophylaxis. J Clin Microbiol 26:1216-8. 125. Kashuba, A. D., A. N. Nafziger, G. L. Dr usano, and J. S. Bertino, Jr. 1999. Optimizing aminoglycoside therapy for nosocomial pne umonia caused by gram-negative bacteria. Antimicrob Agents Chemother 43:623-9. 126. Katsube, T., Y. Yamano, and Y. Yano. 2008. Pharmacokinetic-pharmacodynamic modeling and simulation for in vivo bactericidal effect in murine infection model. J Pharm Sci 97:1606-14. 127. Kearns, G. L., S. M. Abdel-Rahman, J. L. Blum er, M. D. Reed, L. P. James, R. F. Jacobs, J. A. Bradley, I. R. Welshman, G. L. Jungbluth, and D. J. Stalker. 2000. Single dose pharmacokinetics of linezolid in infants and children. Pediatr In fect Dis J 19:1178-84. 128. Khair, O. A., J. M. Andrews, D. Honeybourne, G. Jevons, F. Vacheron, and R. Wise. 2001. Lung concentrations of telithromycin after oral dosing. J Antimicrob Chemother 47:837-40. 129. Khunvichai A., H. M., Karmann M., Joliff L., Rand K., Derendorf H. 2001. Presented at the 41st Annual Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, USA, 2001. 130. Khunvichai A., K. M., Hirt M., Joliffe L., Derendorf H. 2003. Presented at the 32nd Annual Meeting of the American College of Clinical Pharmacology, Tampa, USA.
134 131. Klugman, K. P. 2003. Implications for antimic robial prescribing of strategies based on bacterial eradication. Int J Infect Dis 7:S27-31. 132. Koal, T., M. Deters, K. Resch, and V. Kaever. 2006. Quantification of the carbapenem antibiotic ertapenem in human plasma by a validated liquid chromatography-mass spectrometry method. Clin Chim Acta 364:239-45. 133. Kovar, A., T. Dalla Costa, and H. Derendorf. 1997. Comparison of plasma and free tissue levels of ceftriaxone in rats by microdialysis. J Pharm Sci 86:52-6. 134. Kreilgaard, M. 2002. Assessment of cutane ous drug delivery using microdialysis. Adv Drug Deliv Rev 54 Suppl 1:S99-121. 135. Kreilgaard, M., M. J. Kemme, J. Burggraaf, R. C. Schoemaker, and A. F. Cohen. 2001. Influence of a microemulsion vehicle on cu taneous bioequivalence of a lipophilic model drug assessed by microdialysis and pharmacodynamics. Pharm Res 18:593-9. 136. Kunin, C. M., W. A. Craig, M. Korngut h, and R. Monson. 1973. Influence of binding on the pharmacologic activity of antibio tics. Ann N Y Acad Sci 226:214-24. 137. Kuti, J. L., P. K. Dandekar, C. H. Nigh tingale, and D. P. Nicolau. 2003. Use of Monte Carlo simulation to design an optimized pharmacodynamic dosing strategy for meropenem. J Clin Pharmacol 43:1116-23. 138. Kuti, J. L., N. R. Florea, C. H. Nighti ngale, and D. P. Nicolau. 2004. Pharmacodynamics of meropenem and imipenem against Enteroba cteriaceae, Acinetobacter baumannii, and Pseudomonas aeruginosa. Pharmacotherapy 24:8-15. 139. Laethem, T., I. De Lepeleire, J. McCrea, J. Zhang, A. Majumdar, D. Musson, D. Rogers, S. Li, M. Guillaume, A. Parneix-Spake, and P. Deutsch. 2003. Tissue penetration by ertapenem, a parenteral carbapenem admini stered once daily, in suction-induced skin blister fluid in healthy young volunteers. Antimicrob Agents Chemother 47:1439-42. 140. Lang, S. D., G. L. Cameron, and P. R. Mullins. 1981. Anomalous effect of serum on the antimicrobial activity of cefoperazone. Drugs 22 Suppl 1:52-9. 141. Laue, H., T. Valensise, A. Seguin, S. Hawser S. Lociuro, and K. Islam. 2007. Effect of human plasma on the antimicrobial activity of iclaprim in vitro. J Antimicrob Chemother 60:1388-90. 142. Lee, B. L., M. Sachdeva, and H. F. Chambers. 1991. Effect of protein binding of daptomycin on MIC and antibacterial activ ity. Antimicrob Agents Chemother 35:2505-8. 143. Lees, P., and F. Shojaee Aliabadi. 2002. Rational dosing of antimicrobial drugs: animals versus humans. Int J Antimicrob Agents 19:269-84.
135 144. Leggett, J. E., and W. A. Craig. 1989. Enha ncing effect of seru m ultrafiltrate on the activity of cephalosporins agai nst gram-negative bacilli. Antimicrob Agents Chemother 33:35-40. 145. Leveque, N., S. Makki, J. Hadgraft, and P. Humbert. 2004. Comparison of Franz cells and microdialysis for assessing salicylic acid penetration through human skin. Int J Pharm 269:323-8. 146. Levin, B. R., and D. E. Rozen. 2006. Non-inherited antibiotic resistance. Nat Rev Microbiol 4:556-62. 147. Li, R. C. 2000. New pharmacodynamic parameters for antimicrobial agents. Int J Antimicrob Agents 13:229-35. 148. Li, R. C., M. Zhu, and J. J. Schenta g. 1999. Achieving an optimal outcome in the treatment of infections. The role of clin ical pharmacokinetics and pharmacodynamics of antimicrobials. Clin Pharmacokinet 37:1-16. 149. Lin, Y. H., V. C. Wu, I. J. Tsai, Y. L. Ho, J. J. Hwang, Y. K. Tsau, C. Y. Wu, K. D. Wu, and P. R. Hsueh. 2006. High frequency of lin ezolid-associated thrombocytopenia among patients with renal insufficienc y. Int J Antimicrob Agents 28:345-51. 150. Lindemann, U., K. Wilken, H. J. Weigmann, H. Schaefer, W. Sterry, and J. Lademann. 2003. Quantification of the horny layer using ta pe stripping and micr oscopic techniques. J Biomed Opt 8:601-7. 151. Lipsky, B. A., D. G. Armstrong, D. M. Citron, A. D. Tice, D. E. Morgenstern, and M. A. Abramson. 2005. Ertapenem versus piperacillin/tazobact am for diabetic foot infections (SIDESTEP): prospective, randomised, controlled, double-blinded, multicentre trial. Lancet 366:1695-703. 152. Liu, P., and H. Derendorf. 2003. Antimicrobial tissue concentrations. Infect Dis Clin North Am 17:599-613. 153. Liu Q., L. N., Hintze S., Rand K., Dere ndorf H. 2002. Presented at the AAPS Annual Meeting and Exposition, Toronto. 154. Livermore, D. M., M. W. Carter, S. Bagel, B. Wiedemann, F. Baquero, E. Loza, H. P. Endtz, N. van Den Braak, C. J. Fernandes, L. Fernandes, N. Frimodt-Moller, L. S. Rasmussen, H. Giamarellou, E. Giamarellos-Bourboulis, V. Jarlier, J. Nguyen, C. E. Nord, M. J. Struelens, C. Nonhoff, J. Turnidge, J. Bell, R. Zbinden, S. Pfister, L. Mixson, and D. L. Shungu. 2001. In vitro activities of ertapenem (MK-0826) against recent clinical bacteria collected in Europe and Australia. Antimicrob Agents Chemother 45:1860-7.
136 155. Livermore, D. M., K. J. Oakton, M. W. Carter, and M. Warner. 2001. Activity of ertapenem (MK-0826) versus Enterobacter iaceae with potent beta-lactamases. Antimicrob Agents Chemother 45:2831-7. 156. Livermore, D. M., A. M. Sefton, and G. M. Scott. 2003. Properties and potential of ertapenem. J Antimicrob Chemother 52:331-44. 157. Livermore, D. M., M. Warner, S. Mushta q, S. North, and N. Woodford. 2007. In vitro activity of the oxazolidin one RWJ-416457 against linezolid-re sistant and -susceptible staphylococci and enterococci. Antimicrob Agents Chemother 51:1112-4. 158. Lodise, T. P., Jr., D. H. Rhoney, V. H. Tam, P. S. McKinnon, and G. L. Drusano. 2006. Pharmacodynamic profiling of cefepime in plas ma and cerebrospinal fluid of hospitalized patients with external ventriculostomie s. Diagn Microbiol Infect Dis 54:223-30. 159. Lodise, T. P., B. M. Lomaestro, and G. L. Drusano. 2006. Application of antimicrobial pharmacodynamic concepts into clinical practice: focus on beta-lactam antibiotics: insights from the Society of Infectious Di seases Pharmacists. Pharmacotherapy 26:132032. 160. Lomaestro, B. M., and G. L. Drusano. 2005. Pharmacodynamic evaluation of extending the administration time of meropenem us ing a Monte Carlo simulation. Antimicrob Agents Chemother 49:461-3. 161. Lonnroth, P., P. A. Jansson, and U. Smith. 1987. A microdialysis method allowing characterization of intercellular water space in humans. Am J Physiol 253:E228-31. 162. Lorentzen, H., F. Kallehave, H. J. Kolmos, U. Knigge, J. Bulow, and F. Gottrup. 1996. Gentamicin concentrations in human subcutaneous tissue. Antimicrob Agents Chemother 40:1785-9. 163. Low, D. E., N. Keller, A. Barth, and R. N. Jones. 2001. Clinical prevalence, antimicrobial susceptibility, and geographic resistance pa tterns of enterococci: results from the SENTRY Antimicrobial Surveillance Program, 1997-1999. Clin Infect Dis 32 Suppl 2:S133-45. 164. Lowdin, E., I. Odenholt, S. Bengtsson, a nd O. Cars. 1996. Pharmacodynamic effects of sub-MICs of benzylpenicillin against Stre ptococcus pyogenes in a newly developed in vitro kinetic model. Antimicrob Agents Chemother 40:2478-82. 165. Lowy, F. D. 2003. Antimicrobial resistance : the example of Staphylococcus aureus. J Clin Invest 111:1265-73. 166. Lucini, V., Grosso, S., Pannacci, M., Scag lione, F. 2006. Presented at the Program and abstracts of the 46th Interscience C onference on Antimicrobial Agents and Chemotherapy, San Francisco, USA.
137 167. MacGowan, A. P. 2004. Elements of design: the knowledge on which we build. Clin Microbiol Inf ect 10:6-11. 168. MacGowan, A. P. 2003. Pharmacokinetic a nd pharmacodynamic profile of linezolid in healthy volunteers and patients with Grampositive infections. J Antimicrob Chemother 51 Suppl 2:ii17-25. 169. MacGowan, A. P., K. E. Bowker, C. M. Tobi n, A. R. Noel, A. M. Lovering, and A. H. Thomson. 2008. Pharmacodynamics of vancom ycin against MRSA: implications of bacterial inoculum for clinical breakpoints, ECCMID, Barcelona. 170. MacGowan, A. P., A. R. Noel, C. A. R ogers, and K. E. Bowker. 2004. Antibacterial effects of amoxicillin-clavulanate against Streptococcus pneumoniae and Haemophilus influenzae strains for which MICs are high, in an in vitro pharmacokinetic model. Antimicrob Agents Chemother 48:2599-603. 171. Marchese, A., G. Balistreri, E. Tonoli, E. A. Debbia, and G. C. Schito. 2000. Heterogeneous vancomycin resistance in me thicillin-resistant Staphylococcus aureus strains isolated in a large Italian hospital. J Clin Mi crobiol 38:866-9. 172. Marchese, A., L. Gualco, A. M. Schito, E. A. Debbia, and G. C. Schito. 2004. In vitro activity of ertapenem against selected respiratory pat hogens. J Antimicrob Chemother 54:944-51. 173. McCleverty, D., R. Lyons, and B. Henry. 2006. Microdialysis sampli ng and the clinical determination of topical dermal bi oequivalence. Int J Pharm 308:1-7. 174. McKinnon, P. S., and S. L. Davis. 2004. Pharmacokinetic and pharmacodynamic issues in the treatment of bacterial infectious dis eases. Eur J Clin Microbiol Infect Dis 23:27188. 175. Meibohm, B., and H. Derendorf. 1997. Basic concepts of pharmacokinetic/pharmacodynamic (PK/PD) mo delling. Int J Clin Pharmacol Ther 35:401-13. 176. Meibohm, B., and H. Derendorf. 2002. Phar macokinetic/pharmacodynamic studies in drug product development. J Pharm Sci 91:18-31. 177. Merrikin, D. J., J. Briant, and G. N. Rolinson. 1983. Effect of protein binding on antibiotic activity in vivo. J Antimicrob Chemother 11:233-8. 178. Miller, R., W. Ewy, B. W. Corrigan, D. Ou ellet, D. Hermann, K. G. Kowalski, P. Lockwood, J. R. Koup, S. Donevan, A. El-Katta n, C. S. Li, J. L. Werth, D. E. Feltner, and R. L. Lalonde. 2005. How modeling and simulation have enhanced decision making in new drug development. J Phar macokinet Pharmacodyn 32:185-97.
138 179. Mitchell, J. A., P. Akarasereenont, C. Th iemermann, R. J. Flower, and J. R. Vane. 1993. Selectivity of nonsteroidal antiinflammatory drugs as inhibitors of constitutive and inducible cyclooxygenase. Proc Natl Acad Sci U S A 90:11693-7. 180. Moore, R. D., P. S. Lietman, and C. R. Smith. 1987. Clinical response to aminoglycoside therapy: importance of the ratio of peak concentration to minimal inhibitory concentration. J Infect Dis 155:93-9. 181. Moribe, K., and K. Maruyama. 2002. Ph armaceutical design of the liposomal antimicrobial agents for infectious disease. Curr Pharm Des 8:441-54. 182. Mouton, J. W., M. N. Dudley, O. Cars H. Derendorf, and G. L. Drusano. 2002. Standardization of pharmacokinetic/pharmacodynamic (PK/PD) terminology for antiinfective drugs. Int J Antimicrob Agents 19:355-8. 183. Mouton, J. W., M. N. Dudley, O. Cars H. Derendorf, and G. L. Drusano. 2005. Standardization of pharmacokinetic/pharmacodynamic (PK/PD) terminology for antiinfective drugs: an update. J Antimicrob Chemother 55:601-7. 184. Mouton, J. W., A. Schmitt-Hoffmann, S. Shap iro, N. Nashed, and N. C. Punt. 2004. Use of Monte Carlo simulations to select ther apeutic doses and provisional breakpoints of BAL9141. Antimicrob Agents Chemother 48:1713-8. 185. Mouton, J. W., U. Theuretzb acher, W. A. Craig, P. M. Tulkens, H. Derendorf, and O. Cars. 2008. Tissue concentrations: do we ever learn? J Antimicrob Chemother 61:235-7. 186. Mouton, J. W., D. J. Touzw, A. M. Ho rrevorts, and A. A. Vinks. 2000. Comparative pharmacokinetics of the carbapenems: clinical implications. Clin Pharmacokinet 39:185201. 187. Mouton, J. W., and A. A. Vinks. 2005. Ph armacokinetic/pharmacodynamic modelling of antibacterials in vitro and in vivo using bact erial growth and kill kinetics: the minimum inhibitory concentration versus stationary concentration. Clin Pharmacokinet 44:201-10. 188. Mueller, M., A. de la Pena, and H. De rendorf. 2004. Issues in pharmacokinetics and pharmacodynamics of anti-infective agents: kill curves versus MIC. Antimicrob Agents Chemother 48:369-77. 189. Muller, M., M. Brunner, R. Schmid, E. M. Pu tz, A. Schmiedberger, I. Wallner, and H. G. Eichler. 1998. Comparison of three different e xperimental methods for the assessment of peripheral compartment pharmacokinetic s in humans. Life Sci 62:PL227-34. 190. Muller, M., O. Haag, T. Burgdorff, A. Georgopoulos, W. Weninger, B. Jansen, G. Stanek, H. Pehamberger, E. Agneter, a nd H. G. Eichler. 1996. Characterization of peripheral-compartment kinetics of antibio tics by in vivo microdialysis in humans. Antimicrob Agents Chemother 40:2703-9.
139 191. Muller, M., H. Mascher, C. Kikuta, S. Sc hafer, M. Brunner, G. Dorner, and H. G. Eichler. 1997. Diclofenac concentrations in defined tissue layers after topical administration. Clin Pharmacol Ther 62:293-9. 192. Muller, M., R. Schmid, A. Georgopoulos, A. Buxbaum, C. Wasicek, and H. G. Eichler. 1995. Application of microdialysis to clini cal pharmacokinetics in humans. Clin Pharmacol Ther 57:371-80. 193. Muller, M., H. Stass, M. Brunner, J. G. Moller, E. Lackner, and H. G. Eichler. 1999. Penetration of moxifloxacin in to peripheral compartments in humans. Antimicrob Agents Chemother 43:2345-9. 194. Neuhauser, M. M., J. L. Prause, L. H. Danziger, and S. L. Pendland. 2003. In vitro bactericidal activities of ABT-773 against ermB strains of Streptococcus pneumoniae. Antimicrob Agents Chemother 47:1132-4. 195. Nichols, R. L., posting date. Preventing Surgic al Site Infections: A Surgeon's Perspective. [Online.] 196. Nicolau, D. P. 2003. Optimizing outcome s with antimicrobial therapy through pharmacodynamic profiling. J Infect Chemother 9:292-6. 197. Nicolau, D. P., and P. G. Ambrose. 2001. Pharmacodynamic profiling of levofloxacin and gatifloxacin using Monte Carlo simula tion for community-acquired isolates of Streptococcus pneumoniae. Am J Med 111:13-18; 36-38. 198. Nielsen, E. I., A. Viberg, E. Lowdin, O. Cars, M. O. Karlsson, and M. Sandstrom. 2006. A Semi-Mechanistic Pharmacokinetic/Pharmacodynamic Model for the Assessment of Activity of Antibacterial Agents from Time-Kill Curve Experiments. Antimicrob Agents Chemother 51:128-36. 199. Nix, D. E., A. K. Majumdar, and M. J. DiNubile. 2004. Pharmacokinetics and pharmacodynamics of ertapenem: an overview for clinicians. J Antimicrob Chemother 53 Suppl 2:ii23-8. 200. Nix, D. E., K. R. Matthias, and E. C. Ferguson. 2004. Effect of ertapenem protein binding on killing of bacteria. Antimicr ob Agents Chemother 48:3419-24. 201. Noel, A. R., K. E. Bowker, and A. P. Macgowan. 2005. Pharmacodynamics of moxifloxacin against anaerobes studied in an in vitro pharmacokine tic model. Antimicrob Agents Chemother 49:4234-9. 202. Nolting, A., T. Dalla Costa, K. H. Rand, and H. Derendorf. 1996. Pharmacokineticpharmacodynamic modeling of the antibiotic eff ect of piperacillin in vitro. Pharm Res 13:91-6.
140 203. Noreddin, A. M., D. J. Hoban, and G. G. Zhanel. 2005. Comparison of gatifloxacin and levofloxacin administered at various dosing regimens to hospitalised patients with community-acquired pneumonia: pharmacodynamic target attainment study using North American surveillance data for Streptococcus pneumoniae. Int J Antimicrob Agents 26:120-5. 204. Nowak, G., J. Ungerstedt, J. Wernerma n, U. Ungerstedt, an d B. G. Ericzon. 2002. Clinical experience in conti nuous graft monitoring with microdialysis early after liver transplantation. Br J Surg 89:1169-75. 205. Odenholt, I., and O. Cars. 2006. Pharmac odynamics of moxifloxacin and levofloxacin against Streptococcus pneumoniae, Staphyloc occus aureus, Klebsiella pneumoniae and Escherichia coli: simulation of human plasma concentrations after intravenous dosage in an in vitro kinetic model. J Antimicrob Chemother 58:960-5. 206. Odenholt, I., O. Cars, and E. Lowdin. 2004. Pharmacodynamic studies of amoxicillin against Streptococcus pneumoniae: comparis on of a new pharmacokinetically enhanced formulation (2000 mg twice daily) with standard dosage regimens. J Antimicrob Chemother 54:1062-6. 207. Olson, R. J., and J. B. Justice, Jr. 199 3. Quantitative microdial ysis under transient conditions. Anal Chem 65:1017-22. 208. Palavecino, E. 2004. Community-acquired meth icillin-resistant Staphylococcus aureus infections. Clin Lab Med 24:403-18. 209. Palumbi, S. R. 2001. Humans as the worl d's greatest evolutionary force. Science 293:1786-90. 210. Pearson, R. D., R. T. Steigbigel, H. T. Davis, and S. W. Chapman. 1980. Method of reliable determination of minimal lethal an tibiotic concentrations. Antimicrob Agents Chemother 18:699-708. 211. Perry, T. R., and J. J. Schentag. 2001. Clinical use of ceftriaxone: a pharmacokineticpharmacodynamic perspective on the impact of minimum inhibitory concentration and serum protein binding. Clin Pharmacokinet 40:685-94. 212. Popick, A. C., W. G. Crouthamel, and I. Bekersky. 1987. Plasma protein binding of ceftriaxone. Xenobiotica 17:1139-45. 213. Presant, C. A., W. Wolf, V. Waluch, C. Wiseman, P. Kennedy, D. Blayney, and R. R. Brechner. 1994. Association of intratumor al pharmacokinetics of fluorouracil with clinical response. Lancet 343:1184-7.
141 214. Preston, S. L., G. L. Drusano, A. L. Berman, C. L. Fowler, A. T. Chow, B. Dornseif, V. Reichl, J. Natarajan, and M. Corrado. 1998. Pharmacodynamics of levofloxacin: a new paradigm for early clinic al trials. Jama 279:125-9. 215. Radermacher, J., D. Jentsch, M. A. Scholl, T. Lustinetz, and J. C. Frolich. 1991. Diclofenac concentrations in synovial fluid and plasma after cuta neous application in inflammatory and degenerative joint di sease. Br J Clin Pharmacol 31:537-41. 216. Regoes, R. R., C. Wiuff, R. M. Zappala, K. N. Garner, F. Baquero, and B. R. Levin. 2004. Pharmacodynamic functions: a multiparameter approach to the design of antibiotic treatment regimens. Antimicrob Agents Chemother 48:3670-6. 217. Riess, W., K. Schmid, L. Botta, K. Kobayash i, J. Moppert, W. Schneider, A. Sioufi, A. Strusberg, and M. Tomasi. 1986. [The perc utaneous absorption of diclofenac]. Arzneimittelforschung 36:1092-6. 218. Riond, J. L., and J. E. Riviere. 1989. Doxycyc line binding to plasma albumin of several species. J Vet Pharmacol Ther 12:253-60. 219. Riviere, J. E., and P. L. Williams. 1992. Ph armacokinetic implications of changing blood flow in skin. J Pharm Sci 81:601-2. 220. Rodvold, K. A. 2001. Pharmacodynamics of an tiinfective therapy: taking what we know to the patient's bedside. Pharmacotherapy 21:319-330. 221. Rollins, D. 1990. Clinical Pharmacokinetics, Eighteenth Edition ed. Mack Publishing Company. 222. Ryan, D. M. 1993. Pharmacokinetics of antibiotics in natural and experimental superficial compartments in animals and humans. J Antimicrob Chemother 31 Suppl D:116. 223. Rybak, M. J. 2006. Pharmacodynamics: relation to antimicrobial resistance. Am J Infect Control 34:38-45; 64-73. 224. Scaglione, F., G. Demartini, M. M. Ar cidiacono, S. Dugnani, and F. Fraschini. 1998. Influence of protein binding on the pharm acodynamics of ceftazidime or ceftriaxone against gram-positive and gram-negative bacteria in an in vitro infection model. J Chemother 10:29-34. 225. Scaglione, F., and L. Paraboni. 2006. Infl uence of pharmacokinetics/pharmacodynamics of antibacterials in their dosing regimen se lection. Expert Rev Anti Infect Ther 4:479-90. 226. Schaper, K. J., S. Schubert, and A. Da lhoff. 2005. Kinetics and quantification of antibacterial effects of beta-lactams, macrolides, and quinolones against gram-positive and gram-negative RTI pathogens. Infection 33:3-14.
142 227. Schito, G. C. 2006. The importance of the development of antibiotic resistance in Staphylococcus aureus. Clin Micr obiol Infect 12 Suppl 1:3-8. 228. Schmidt, S., A. Barbour, M. Sahre, K. H. Rand, and H. Derendorf. 2008. PK/PD: new insights for antibacterial and antiviral applications. Curr Opin Pharmacol. 229. Schmidt, S., E. Schuck, V. Kumar, O. Bur khardt, and H. Derendorf. 2007. Integration of pharmacokinetic/pharmacodynamic modeling and simulation in the development of new anti-infective agents minimum inhibitory c oncentration versus time-kill curves. Expert Opin. Drug Discov. 2:849-860. 230. Schuck, E. L., A. Dalhoff, H. Stass, and H. Derendorf. 2005. Pharmacokinetic/pharmacodynamic (PK/PD) eval uation of a once-daily treatment using ciprofloxacin in an extended-rel ease dosage form. Infection 33:22-8. 231. Schuck, E. L., and H. Derendorf. 2005. Pharmacokinetic/pharmacodynamic evaluation of anti-infective agents Expert Rev Anti Infect Ther 3:361-73. 232. Sevillano, D., M. J. Gimenez, L. Alou, L. A guilar, F. Cafini, M. Torrico, N. Gonzalez, O. Echeverria, P. Coronel, and J. Prieto. 2007. Effects of human albumin and serum on the in vitro bactericidal activity of cefditoren against penicillin-resistant Streptococcus pneumoniae. J Antimicrob Chemother 60:156-8. 233. Shefet, D., E. Robenshtok, M. Paul, and L. Leibovici. 2005. Empirical atypical coverage for inpatients with community-acquired pne umonia: systematic review of randomized controlled trials. Arch Intern Med 165:1992-2000. 234. Singh, P., and M. S. Roberts. 1994. Effects of vasoconstriction on dermal pharmacokinetics and local tissue distribu tion of compounds. J Pharm Sci 83:783-91. 235. Singh, P., and M. S. Roberts. 1994. Skin perm eability and local tissu e concentrations of nonsteroidal anti-inflammatory drugs after topical application. J Pharmacol Exp Ther 268:144-51. 236. Smith, H. 1998. What happens to bacterial pa thogens in vivo? Tre nds Microbiol 6:23943. 237. Smith, P. F., B. Tsuji, B. M. Booker, A. Forrest, S. Bajic, P. Kelchlin, S. M. Bhavnani, R. N. Jones, and P. G. Ambrose. 2006. Pharmacodynamics of cefprozil against Haemophilus influenzae in an in vitro pharmacodynamic model. Diagn Microbiol Infect Dis 56:379386. 238. Song, J. H. 2003. Introduction: the goals of an timicrobial therapy. Int J Infect Dis 7:1-4. 239. Soriano, A., O. Miro, and J. Mensa. 2005. Mitochondrial toxicity associated with linezolid. N Engl J Med 353:2305-6.
143 240. Stagni, G., D. O'Donnell, Y. J. Liu, D. L. Kellogg, T. Morgan, and A. M. Shepherd. 2000. Intradermal microdialysis: kinetics of iontophoretically delivered propranolol in forearm dermis. J Control Release 63:331-9. 241. Stahl, M., R. Bouw, A. Jackson, and V. Pay. 2002. Human microdialysis. Curr Pharm Biotechnol 3:165-78. 242. Stahle, L., P. Arner, and U. Ungers tedt. 1991. Drug distribution studies with microdialysis. III: Extracellular concentration of caffeine in adipose tissue in man. Life Sci 49:1853-8. 243. Stalker, D. J., G. L. Jungbluth, N. K. H opkins, and D. H. Batts. 2003. Pharmacokinetics and tolerance of singleand multiple-dose oral or intravenous linezolid, an oxazolidinone antibiotic, in healt hy volunteers. J Antimicrob Chemother 51:1239-46. 244. Stoeckel, K. 1981. Pharmacokinetics of Ro cephin, a highly active new cephalosporin with an exceptionally long biologi cal half-life. Chemotherapy 27:42-6. 245. Stoeckel, K., P. J. McNamara, R. Brandt, H. Plozza-Nottebrock, and W. H. Ziegler. 1981. Effects of concentration-dependent plasma pr otein binding on ceftriaxone kinetics. Clin Pharmacol Ther 29:650-7. 246. Suh, B., W. A. Craig, A. C. England, and R. L. Elliott. 1981. Effect of free fatty acids on protein binding of antimicrobial agents. J Infect Dis 143:609-16. 247. Svensson, C. K., M. N. Woodruff, J. G. Baxter, and D. Lalka. 1986. Free drug concentration monitoring in clinical pract ice. Rationale and cu rrent status. Clin Pharmacokinet 11:450-69. 248. Tam, V. H., P. S. McKinnon, R. L. Akins, G. L. Drusano, and M. J. Rybak. 2003. Pharmacokinetics and pharmacodynamics of cefep ime in patients with various degrees of renal function. Antimicrob Agents Chemother 47:1853-61. 249. Tam, V. H., A. N. Schilling, K. Poole, and M. Nikolaou. 2007. Mathematical modelling response of Pseudomonas aeruginosa to mer openem. J Antimicrob Chemother 60:1302-9. 250. Tegeder, I., L. Brautigam, M. Podda, S. Me ier, R. Kaufmann, G. Geisslinger, and M. Grundmann-Kollmann. 2002. Time course of 8methoxypsoralen concentrations in skin and plasma after topical (bath and cream) a nd oral administration of 8-methoxypsoralen. Clin Pharmacol Ther 71:153-61. 251. Theuretzbacher, U. 2007. Tissue penetration of antibacterial agents: how should this be incorporated into pharmacodynamic an alyses? Curr Opin Pharmacol 7:498-504.
144 252. Thomas, J. K., A. Forrest, S. M. Bhavnani, J. M. Hyatt, A. Cheng, C. H. Ballow, and J. J. Schentag. 1998. Pharmacodynamic evaluation of factors associated with the development of bacterial resistance in acutely ill patients during therapy. Antimicrob Agents Chemother 42:521-7. 253. Thorsen, K., A. O. Kristoffersson, U. H. Lerner, and R. P. Lorentzon. 1996. In situ microdialysis in bone tissue. Stimulation of prostaglandin E2 release by weight-bearing mechanical loading. J Clin Invest 98:2446-9. 254. Tomaselli, F., A. Maier, V. Matzi, F. M. Smolle-Juttner, and P. Dittrich. 2004. Penetration of meropenem into pneumonic human lung tissue as measured by in vivo microdialysis. Antimicrob Agents Chemother 48:2228-32. 255. Toutain, P. L., and A. Bousquet-Me lou. 2002. Free drug fraction vs free drug concentration: a matter of frequent c onfusion. J Vet Pharmacol Ther 25:460-3. 256. Toutain, P. L., J. R. del Castillo, an d A. Bousquet-Melou. 2002. The pharmacokineticpharmacodynamic approach to a rational dosag e regimen for antibiotics. Res Vet Sci 73:105-14. 257. Travers, K., and M. Barza. 2002. Morbidit y of infections caused by antimicrobialresistant bacteria. Clin In fect Dis 34 Suppl 3:S131-4. 258. Treyaprasert, W., S. Schmidt, K. H. Rand, U. Suvanakoot, and H. Derendorf. 2006. Pharmacokinetic/pharmacodynamic modeling of in vitro activity of azithromycin against four different bacterial strains. Int J Antimicrob Agents. 29:263-70 259. Tsiodras, S., H. S. Gold, G. Sakoulas G. M. Eliopoulos, C. Wennersten, L. Venkataraman, R. C. Moellering, and M. J. Ferraro. 2001. Linezolid resistance in a clinical isolate of Staphyloc occus aureus. Lancet 358:207-8. 260. Ungerstedt, U., and C. Pycock. 1974. Functional correlates of dopamine neurotransmission. Bull Schwei z Akad Med Wiss 30:44-55. 261. Van der Auwera, P., and J. Klastersky. 1990. Study of the influence of protein binding on serum bactericidal titres and killing rates in volunteers receiving ceftazidime, cefotaxime and ceftriaxone. J Hosp Infect 15 Suppl A:23-34. 262. Veronese, F. M., R. Bevilacqua, E. Bo ccu, and C. A. Benassi. 1977. Drug-protein interaction: the binding of cephalospor ins to albumins. Farmaco [Sci] 32:303-10. 263. Vilar-Compte, D., R. Roldan-Marin, C. Robles-Vidal, and P. Volkow. 2006. Surgical site infection (SSI) rates among patients who underw ent mastectomy after the introduction of SSI prevention policies. Infect Control Hosp Epidemiol 27:829-34.
145 264. Walker, K. J., A. J. Larsson, R. A. Zabi nski, and J. C. Rotschafer. 1994. Evaluation of antimicrobial activities of clarithromycin and 14-hydroxyclarithro mycin against three strains of Haemophilus influenzae by using an in vitro pharmacodynamic model. Antimicrob Agents Chemother 38:2003-7. 265. Weber, D. J. 2006. Collateral damage and what the future might hold. The need to balance prudent antibiotic utilization and stew ardship with effective patient management. Int J Infect Dis 10:17-24. 266. Wennberg, A. M., O. Larko, P. Lonnroth, G. Larson, and A. L. Krogstad. 2000. Deltaaminolevulinic acid in superficial basal ce ll carcinomas and normal skin-a microdialysis and perfusion study. Clin Exp Dermatol 25:317-22. 267. Weston, J. S. 2002. Treatment of gram-positive infections: past, present, and future. Crit Care Nurs Clin North Am 14:17-29. 268. White, R. 2001. What in vitro models of infection can and cannot do. Pharmacotherapy 21:292-301. 269. White, R. L., M. B. Kays, T. A. Armstrong, and L. V. Friedrich. 1990. Effects of supraphysiologic temperature and broth d ilution on serum protein binding. Antimicrob Agents Chemother 34:1257-8. 270. WHO January, 2002, posting date. Antim icrobial Resistance. [Online.] 271. Winokur, P. L., R. Canton, J. M. Casellas, and N. Legakis. 2001. Variations in the prevalence of strains expressing an exte nded-spectrum beta-lactamase phenotype and characterization of isolates from Europe, th e Americas, and the Western Pacific region. Clin Infect Dis 32 Suppl 2:S94-103. 272. Wispelwey, B. 2005. Clinical implications of pharmacokinetics and pharmacodynamics of fluoroquinolones. Clin Infect Dis 41:127-35. 273. Wittau, M., E. Wagner, V. Kaever, T. Koal, D. Henne-Bruns, and R. Isenmann. 2006. Intraabdominal tissue concentration of er tapenem. J Antimicrob Chemother 57:312-6. 274. Yamaguchi, K., and A. Ohno. 1992. [Consideratio ns on MRSA infections in relation to modern chemotherapy]. Nippon Rinsho 50:923-31. 275. Yano, Y., T. Oguma, H. Nagata, and S. Sa saki. 1998. Application of logistic growth model to pharmacodynamic analysis of in vitro bactericidal kinetics. J Pharm Sci 87:1177-83. 276. Yassin, H. M., and L. L. Dever. 2001. Telithro mycin: a new ketolide antimicrobial for treatment of respiratory tract infections Expert Opin Investig Drugs 10:353-67.
146 277. Yuk, J. H., C. H. Nightingale, and R. Qu intiliani. 1989. Clinical pharmacokinetics of ceftriaxone. Clin Ph armacokinet 17:223-35. 278. Zeitlinger, M., M. Muller, and C. Joukhada r. 2005. Lung microdialysis--a powerful tool for the determination of exogenous and endoge nous compounds in the lower respiratory tract (mini-review). Aaps J 7:E600-8. 279. Zeitlinger, M., R. Sauermann, M. Fille, J. Hausdorfer, I. Leitner, and M. Muller. 2008. Plasma protein binding of fluoroquinolones affects antimicrobial activity. J Antimicrob Chemother 61:561-7. 280. Zeitlinger, M. A., R. Sauermann, F. Traunmuller, A. Georgopoulos, M. Muller, and C. Joukhadar. 2004. Impact of plasma protein binding on antimicrobial activity using timekilling curves. J Antimicrob Chemother 54:876-80. 281. Zhanel, G. G., C. Johanson, J. M. Embil, A. Noreddin, A. Gin, L. Vercaigne, and D. J. Hoban. 2005. Ertapenem: review of a new carba penem. Expert Rev Anti Infect Ther 3:23-39. 282. Zurenko, G. E., B. H. Yagi, R. D. Schaadt, J. W. Allison, J. O. Kilburn, S. E. Glickman, D. K. Hutchinson, M. R. Barbachyn, and S. J. Brickner. 1996. In vitro activities of U100592 and U-100766, novel oxazolidinone antibact erial agents. Antimicrob Agents Chemother 40:839-45.
147 BIOGRAPHICAL SKETCH Stephan Schm idt was born in 1978, in Sonnebe rg, Germany. The elder of two sons, he grew up mostly in Sonneberg and received hi s high school degree from the Arnold-Gymnasium in Neustadt bei Coburg, Germany, in 1998. He ear ned his B.S. in pharmaceutical sciences from the Friedrich-Alexander University in Erla ngen, Germany, in 2004. On completion of his practical pharmaceutical training year, he was gran ted his license to practice as a pharmacist in Germany in 2005. In May 2005, he joined the Ph.D. program at the University of Florida, College of Pharmacy, in the Department of Pharmaceutics. Under the supervision of Dr. Hartmut Derendorf, distinguished professor and ch air, he focused his research work on the pharmacokinetics and pharmacodynamics of oxa zolidinones and beta-lactams. During his time at the University of Florida, he has been an active member of the American Association of Pharmaceutical Scientists (AAPS), American Society for Microbiology (ASM), American College of Clinical Pharmacology (ACCP), St udent Outreach Committee (SOC) of ACCP and received multiple national and international awards and travel grants. Upon completion of his Ph.D. pr ogram, Stephan will join Prof Meindert Danhofs group at the Leiden University in the Netherlands fo r a 3-year post-doctoral fellowship under the supervision of Dr. Oscar Della Pasqua.