bacterias anaerobias0f

Embed Size (px)

Citation preview

  • 7/21/2019 bacterias anaerobias0f

    1/16

    In vitropharmacodynamic models to determine the effectof antibacterial drugs

    Julia Gloede1, Christian Scheerans1,2, Hartmut Derendorf3 and Charlotte Kloft1,2*

    1Department of Clinical Pharmacy, Institute of Pharmacy, Martin-Luther-Universitaet Halle-Wittenberg, Halle, Germany;2Department of Clinical Pharmacy, Institute of Pharmacy, Freie Universitaet Berlin, Berlin, Germany;

    3Department of Pharmaceutics, College of Pharmacy, University of Florida, Gainesville, FL, USA

    *Corresponding author. Department of Clinical Pharmacy, Institute of Pharmacy, Martin-Luther-Universitaet Halle-Wittenberg, Wolfgang-Langenbeck-

    Str. 4, 06120 Halle, Germany. Tel: 49-345-5525190, Fax: 49-345-5527257; E-mail: [email protected]

    In vitro pharmacodynamic (PD) models are used to obtain useful quantitative information on the effect ofeither single drugs or drug combinations against bacteria. This review provides an overview of in vitro PDmodels and their experimental implementation. Models are categorized on the basis of whether the drug

    concentration remains constant or changes and whether there is a loss of bacteria from the system. Furthersubdifferentiation is based on whether bacterial loss involves dilution of the medium or is associated withdialysis or diffusion. For comprehension of the underlying principles, experimental settings are simplified andschematically illustrated, including the simulations of various in vivo routes of administration. The differentmodel types are categorized and their (dis)advantages discussed. The application of in vitro models tospecial organs, infections and pathogens is comprehensively presented. Finally, the relevance and perspectivesofin vitroinvestigations in drug discovery and clinical research are elucidated and discussed.

    Keywords:in vitro models, antibiotics, dilution models, dialysis/diffusion models, static models

    Introduction

    The dosing regimens of antibiotics are often not optimal and thedoseresponse relationships not well known.1 One importantreason is that in the patient the pure antibiotic effect, i.e. thepharmacodynamic (PD) characteristics,2 cannot clearly be separ-ated from other factors determining the response to the antibac-terial treatment. The effect also has to be regarded along withpharmacokinetic (PK) properties,2 such as the ability of thedrug to reach its target. Thus, the PK, and the PD,2 are character-istics of an antibacterial agent and should be considered in thedevelopment and prediction of the efficacy of the antibacterialtherapy. By linking the concentrationtime course (at the siteof action) to the drug effect (PK/PD), various dosing regimensfor different pathogens can be investigated in silico, enablingthe identification of potentially effective dosing regimens.

    However, there is no standardized procedure for PK/PD evalu-ation for antibiotics, although the European Medicines Agency(EMEA)1 and the FDA3 clearly recommend these investigationsfor new compounds.

    For characterizing the PD of an antibiotic, bacterial growthand death under antibiotic exposure have to be investigated.Since these are difficult to measure in human tissue, animaland in vitro models have been developed. Animal modelsprovide similar growing conditions for bacteria, closely imitatingthe characteristics of a human infection, and the endpoint of an

    infection is clearly defined (cure or death) and comparable tothat in humans.4,5 A significant disadvantage of animal models

    is differences in the PK,1,5

    e.g. in metabolism, which limit ornecessitate sophisticated scaling methods for transferring datafrom animals to humans.5

    In contrast, in vitro models can mimic human PK1 and arethus better suited for the investigation of antibiotic activity.6 Inaddition, they allow resistance analyses,7,8 determination oftime kill behaviour, and the identification and optimizationof PK/PD indices and breakpoints.913 Although a large numberof models have been developed, in practice they are all variantsof 10 basic experimental set-ups. Rather than discuss all themodels reported in the literature, this article provides a general-ized overview of the most frequently used and newly developedin vitroPD models. The historical development ofin vitro modelshas been already reviewed by Grasso until 198514 and

    others.1517

    Grasso divided in vitro models according to theirworking principle into two basic groups: (i) models based ondilution; and (ii) those based on diffusion or dialysis. MacGowanet al.18,19 described the information and conclusions obtainedfromin vitromodels. The impact ofin vitromodels has been dis-cussed by Li and Zhu,17 and others.16 PK modelling of in vitromodels has been basically described by Blaser,20 Rowe and Mor-ozowich,21 and Firsovet al.;15 detailed mathematical descriptionsfor the interpretation of PK/PD analyses and PK/PD modellingcan be found in the work of Derendorf and Meibohm,22 Czock

    # The Author 2009. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.For Permissions, please e-mail: [email protected]

    J Antimicrob Chemother2010; 65: 186201doi:10.1093/jac/dkp434 Advance publication 21 December 2009

    186

  • 7/21/2019 bacterias anaerobias0f

    2/16

    and Keller,23 and others.2437 PK/PD parameters for antibiotics,the PK/PD indices, are well defined in the literature fromMouton et al.10,11

    FundamentalsCharacteristics of in vitro models

    The two main characteristics of in vitro PD models are drugexposure and bacterial concentration. The literature does notcomprise a uniform and complete definition of these maincharacteristics. Hence, we characterize these terms as follows:constant drug exposure is achieved by not replacing or changingthe medium, while changing drug concentrations are obtained insystems with flowing medium. Consequently, we will focus onthe termsconstantand changingto describe the drug exposure.The bacterial concentration represents the magnitude of the PDeffect. A loss of bacteria due to the experimental setting, whichis observed in some models, may therefore have a substantial

    influence on the results. Thus, to define the loss of bacteria inin vitro models we suggest the terms open and closed:38 openmodels allow the exchange of bacteria with the environment;and closed models have no bacterial exchange. As a result, anopen model always has a flowing medium (i.e. changing drugexposure), whereas closed models can have an unchanging orflowing medium (i.e. constant or changing drug exposure).

    Classification

    As a consequence of these definitions, allin vitromodels for anti-biotics can be classified according to the change of the drug con-centration and whether or not there is bacterial loss (Table 1):

    I models with aconstantdrug concentration andnobacterialloss;

    II models with changing drug concentration and bacterialloss; and

    III models withchangingdrug concentration and no bacterialloss.

    The first models can be more accurately named static in vitromodels (Table 1, No. I). This term is already advised in the litera-ture of microbiology and biotechnology for models with a con-stant environment, i.e. constant antibiotic exposure, withunchanged medium;39,40 thus, no in- and outflow of mediumoccurs in these systems.41,42 These models have been usedextensively;28,4348 however, these models had either not beennamed before or have been described as models investigatingconstant concentrations of antibiotics.42

    The other two groups of models are known asdynamic in vitromodels and are further differentiated on the basis of whether ornot bacterial loss occurs (Table 1, Nos II and III). Usually, bac-terial loss38 (No. II) is not intended and causes bias, which canbe corrected,40,49,50 whereas the dilution of toxic waste cannotbe considered. To avoid bacterial loss, appropriate technicalarrangements have to be carried out (No. III), e.g. by a mem-brane or filter system.38 Models Nos II and III can be subclassi-fied depending on whether the mechanism of drug loss involvesdilution (Nos II and IIIa) or dialysis/diffusion (No. IIIb). Dilutionmodels with bacterial loss (No. II) work by stepwise substitutionor continuous dilution of medium. Dilution models without bac-terial loss (No. IIIa) operate by stepwise or continuous dilution, orstepwise substitution of medium through a filter system. In thespecial case of dilution models without bacterial loss (No. IIIa),medium is added, with a resulting increase in volume, i.e. noloss of bacteria, although their concentration is reduced due todilution (see the Stepwise simple dilution section below andTable 2). Dialysis/diffusion models can be further classified bythe type of membrane used (eitherartificial or natural).

    Our review shows that the current classification14 (dilution ordialysis models) does not encompass all models (e.g. intracellu-

    lar models), although they could still be integrated into theclassification scheme. We adapted and revised the existingclassification, and focused on in vitro models, which mimic PKprofiles in plasma and other biological matrices, and thus allow

    Table 1. Revised classification ofin vitromodels

    Drug concentration

    Bacterial loss

    yes (open systems) no (closed systems)

    Constant I static models

    Changing II dynamic dilution models

    via stepwise substitution (without filters) via continuous simple dilutionb (without filters)

    IIIa dynamic dilution models

    via stepwise simple dilutiona

    via stepwise substitution (with filters)

    via continuous dilution

    W without outleta

    W with filtersb

    IIIb dynamic dialysis/diffusion modelsb

    with artificial membranes

    with natural membranes

    aWith bacterial dilution.bMulticompartments possible.

    Review

    187

    JA

  • 7/21/2019 bacterias anaerobias0f

    3/16

  • 7/21/2019 bacterias anaerobias0f

    4/16

    investigations on the relation/s of concentration- and time-dependent bacterial growth. Specific organs and their specialconditions appear separately in Table 2 (Applications). Thenew, extended classification (see Table 1) is based on both men-tioned main characteristics of in vitro models, i.e. drug concen-tration and bacterial concentration, whereby all commonlyused models can be categorized.

    Experimental settings of in vitro models

    Anin vitromodel with its main componentthe culture vesselhas to fulfil special requirements (Table 3). Numerous differentinvitromodels for antibiotics have been developed. Not all modelsare applicable for all purposes, and some can mimic onlyselected aspects and conditions, e.g. infections in specific com-partments. The bacterial concentration (as the antibacterialeffect) in thein vitrosystem is monitored over time under differ-ent antibiotic exposures by different methods (Table 4). The bac-terial concentrationtime courses (timekill curves) and derived

    PK/PD indices, such as area under the timekill/bacterial curve,allow for detailed analysis of bacterial growth and death follow-ing antibiotic exposure.18,19,33,34,41,51

    Common working principles

    In staticin vitromodels, bacteria should be suspended homoge-neously in a culture vessel with constant antibiotic exposure inthe medium. All conditions remain the same over the entireobservation period. The bacterial growth without antibiotic canbe limited by nutrition, space, aeration and toxic metabolites.The bacterial concentration changes in the vessel and can bestudied over time.41,42 The working principle of dynamicmodels is more complex. The idea is to simulate the body clear-

    ance or half-life of the antibiotic and is realized in dynamicmodels by changing drug concentrations (Figure 1).41 In dilutionmodels the drug concentration in the culture vessel changes viasubstitution with fresh medium or by simple dilution. Substi-tution means to remove a defined volume from the in vitromodel and supplement the same volume of fresh medium. Inthis case, both flow processes (i.e. in- and outflow) are con-trolled. The volume in the model remains constant all the time.Simple dilution means to add a defined volume of medium tothe culture vessel. Either (i) medium is added to the input andthe outflow is uncontrolled via overflow (or does not exist) or(ii) a pump removes medium from the culture vessel and freshmedium is sucked in from a reservoir by low pressure. In bothcases, the drug concentration in the culture vessel will be

    diluted. The input of medium in dilution models can happen con-tinuously or stepwise, i.e. at intervals.42 Fresh medium is pumpedfrom a reservoir into the culture vessels and from there into thewaste. The input of the drug can mimic bolus, infusion or first-order absorption. The experimental implementation of in vivoadministration routes is explained below.

    Another method for achieving changing drug concentrationsis via drug diffusion across a membrane (dialysis), with the con-centration gradient as the driving force. The dialysis modelsconsist of a central compartment, where the drug initiallyappears after dosing, and a peripheral compartment, with bac-teria. The central compartment and peripheral compartmentM

    odelsforcombinationtherapyc

    synergisticeffects

    staticmodel

    4

    8,153

    dilutionmodelwithstepwise

    substitution(withfilters)

    7

    3

    dilutionmodelwithcontinuo

    usdilution

    withfilters

    1

    54

    withoutfilter

    s

    2

    0,155157

    dialysismodel

    5

    1,117,158160

    Modelsforfungiandanaerobicorg

    anisms

    staticmodel

    1

    61,162

    dilutionmodelwithcontinuo

    usdilution

    withoutfilter

    s

    1

    21

    withfilter

    1

    63

    dilutionmodelwithstepwise

    simpledilution

    7

    2

    dialysismodel

    withartificialmembranes

    1

    64

    withnaturalmembranes

    1

    62,165

    aProductionofbiofilmsbygrowing

    slime-producingbacteria,e.g.

    Staphylococcusaureus47

    andPseudomonasaeruginosa

    135

    onvarioussurfaces.

    bHostcellsgrownuntilastablecultureappears(continuouslayer),

    bacterialsuspensiondirectlyadded,cultureisincuba

    tedandinvestigationsstartwhentheinfectionispositive.

    cIfantibioticswithdifferenthalf-livesaresimultaneouslyinvestigatedindynam

    icmodels,theflowofthemedium,which

    decreasesthedrug,

    hastobeadjustedtotheshortesthalf-

    life41

    or,moreappropriately,thedrugwiththelongerhalf-lifehastobesubs

    titutedtothereservoir.

    Review

    189

    JA

  • 7/21/2019 bacterias anaerobias0f

    5/16

    are separated by a semi-permeable membrane, i.e. permeablefor drug and medium but not for bacteria. Fresh medium is con-tinuously pumped from a reservoir into the central compartmentand then into the waste. Thus, the medium in the peripheralcompartment is continuously renewed by diffusion (from thecentral compartment), while the drug and bacteria can interact,but the bacteria cannot leave this compartment (Figure 1). Thecirculation of medium in the peripheral compartmentas coun-terflow towards the central compartmentcan help to optimizethe diffusion.52,53

    Model developments

    Static models (No. I)

    Static models consist of a closed culture vessel (Figure 2a). Thesevessels are available in a variety of shapes, such as tubes,46,47

    flasks,43,54 cell culture flasks44 or spinner flasks,55 and may bemade of glass43 or polystyrene.47 The first time kill investi-gations in static models were established by Garrett et al.43

    in 1966.

    Dynamic dilution models with bacterial loss (No. II)

    Stepwise substitution (without filters)

    Nishidaet al.56 described a dilution model with stepwise substi-tution of the medium (Figure 2b). In this model, a tube containsthe bacteria in medium. Fresh medium is periodically added andat the same time the same volume of used medium is discarded,leading to a stepwise decline of the drug and a removal of thebacteria.

    Continuous simple dilution (without filters)

    Models with continuous simple dilution reflect the in vivo con-ditions of a drug much more closely than a stepwise decline ofthe drug. The decisive improvement in this field was made by

    Grasso et al.57

    The Grasso model consists of a flask containingthe bacteria (culture vessel), a reservoir and a waste container(Figure 2b). Fresh medium is continuously pumped from thereservoir into the flask and used medium leaves the culturevessel by the pressure of the incoming medium. Drug and bac-terial samples can be taken from the vessel. A magnetic stirrerensures a homogeneous distribution of the drug and bacteria.In the Grasso model, the bacteria are diluted by the incomingmedium and flow out with the outgoing medium, whichdemands a mathematical correction for bacterial counts. Intheory, flow rates that are faster than the bacterial growthrates would lead to a complete loss of bacteria; however, Haag

    et al.58 demonstrated that bacteria might adhere to the vesselwall, forming a biofilm, which protects the bacteria from(out)flow and from antibiotics. Hence, only released bacteriacan be counted in the medium and higher bacterial concen-trations can be found. Nevertheless, due to its simplicity theGrasso model was adapted by several groups.5968 In a similarmodel, air pressure instead of peristaltic pumps has beenused.69 Bergan et al.70 introduced a second peristaltic pumpfor the out-flowing medium. Later, a computer was added tocontrol three pump sets in parallel.71

    Murakawa et al.40 developed a two-compartment modelbased on the Grasso model (Figure 2c). At the beginning, thedrug is administered as a bolus into the first (central) compart-ment containing the bacteria, with the second (peripheral) com-partment remaining drug free. Fresh medium is pumped fromthe reservoir into the first compartment. A second pumpexchanges the medium between both compartments. Bacterialexchange is not prevented and bacteria are eliminated into thewaste; mathematical corrections have been applied.

    Dynamic dilution models without bacterial loss (No. IIIa)

    Stepwise simple dilutionIn a stepwise simple dilution model the medium is not removedfrom the system. Fresh medium is added periodically and thedrug concentration declines over time, in relation to the increasein the volume of the medium (Figure 2d).72 Simultaneously, bac-teria will be diluted; hence, bacterial concentrations have to becorrected.

    Stepwise substitution (with filters)

    Noltinget al.45 developed a model (syringe model) where thedrug concentration is decreased by stepwise substitution, butthe bacterial loss is prevented by a filter. A syringe needle is

    stuck into a cell culture flask containing bacteria and medium(Figure 2e). The needle is connected with a filter unit and asyringe. Used medium is withdrawn at regular intervals fromthe cell culture flask (in contrast to the stepwise simple dilutionmodel) and replaced by fresh medium.45,73

    Another stepwise substitution model was introduced byHaller,65 and comprises a Teflon-coated ultrafiltration unit filledwith medium and bacteria. After adding the drug, air pressureis applied. Thus, medium is continuously eluted and discarded.Fresh medium, however, is replaced at intervals (Figure 2e).65,74

    The elution can also be performed by centrifugation of the fil-tration unit.75

    Table 3. Requirements forin vitro models

    Parameter Rationale Implementation

    Growth medium appropriate growing conditions

    Temperature choice and control to mimicin vivogrowing condition waterbath, incubator

    Mixing quick homogeneous distribution and aeration of bacterial suspension shaking or stirring

    Review

    190

  • 7/21/2019 bacterias anaerobias0f

    6/16

    Table 4. Quantification methods for bacteria

    Method Principle

    Properties

    online

    measurementadirect

    measurementbdetection of live

    cells only

    differentiation

    between live

    and dead cells

    Viable cell counts incubation of bacterial samples on agar,

    followed by counting

    2 2 NA most fre

    avoid

    carry-

    neces

    Turbidimetry measurement of optical density of

    bacteria in medium (correlates with

    bacterial concentration)

    2 2 discusse

    Impedance measurement of impedance of bacterial

    cells (correlates with bacterial

    concentration)

    2 2 2

    Bioluminescence determination of ATP content of

    bacterial cells released (correlates

    with bacterial concentration)

    2 2 2 2

    Microscope determination of bacteria by a phase

    contrast microscopy

    2 2 has to b

    one g

    bacte

    Fluorescence

    quantification

    determination of release of fluorogenic

    substances from a substrate by

    bacterial phosphatases

    2 NA

    RNA profiling quantitative PCR of RNA (correlates with

    bacterial concentration)

    2 2 NA

    NA, not applicable.aWithin 10 min.bIn the culture vessel.cSpecial caution for carry-over effect should be paid for quinolones.

    191

    byguestonApril27,2014 http://jac.oxfordjournals.org/ Downloadedfrom

    http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/http://jac.oxfordjournals.org/
  • 7/21/2019 bacterias anaerobias0f

    7/16

    Continuous dilution

    Continuous dilution models operatingwithout outlet, resulting inincreasing volumes, were described by Sanfilippo and Morvillo,76

    and OGrady and Pennington.77 They mostly reflect only selectedaspects of the in vivo situation. Pumps transport the medium

    from a reservoir into the culture flask (Figure 2d). Since there isno outlet, the volume of the second flask continuously increases,the drug concentration changes and the bacteria will bediluted.76 The increasing volume does not necessarily allowexact exponential decline of the drug concentrations (see alsobladder/bacterial cystitis in Table 2).77

    Continuous dilution without bacterial loss can also be achievedwith filters. The model by Greenwood and Tupper78 consists of avessel separated by a filter membrane in two chambers, withthe bacteria and drug being added to the upper one(Figure 2e).78 However, this model has not been further used.Instead, for the already mentioned decisive Grasso model(dilution, but bacterial loss),57 different modifications have beensuggested to improve the accuracy. Filters are inserted between

    the culture vessel and the waste, and the outlet has movedfromthe side panel tothe bottomof the flask.7981A practical sol-ution to prevent the bacterial loss in the Grasso model was foundby Lowdinet al.55 (Figure 2e). The base of a spinner flask is modi-fied, including an outlet and a perforated metal support, on whicha filter membrane and a pre-filter are adjusted. Above the mem-brane, a magneticstirrer is placed to prevent membrane blockage.Fresh medium is pumped in via one side arm in the spinner flask.The other arm is prepared with a silicon membrane for repeatedsampling.55,82,83

    A multicompartment model based on the Grasso modelwith retention of bacteria was presented by Navashin et al.84

    (Figure 2f). Several vessels are connected in series and thenumber of vessels depends on the number of compartmentsof the underlying mathematical model. The modeloperates as the Grasso model. Bacterial loss is prevented bya special filtration unit, which is placed between each vessel.

    Dynamic dialysis/diffusion models (No. IIIb)

    In the majority of dialysis/diffusion models (for simplification,further called dialysis models), the setting is as follows: freshmedium is pumped from a reservoir into the central compart-ment and then into the waste (except for the model byAl-Asadi et al.85), thereby decreasing the drug concentration.The hallmark of these models is that the drug (and medium)has to diffuse through a membrane to reach the bacteria inthe peripheral compartment. In consequence, bacterial growthand fresh medium flow happen in two compartments. Twodifferent settings of the central and peripheral compartmenthave been employed:adjacent and embedded.

    Dialysis models can be subclassified by the nature of their

    membrane, i.e. models with artificial and natural membranes(Table 5). In this way, the classification presented here includesmodels that previously have not been named dialysis models.At first glance this might seem unusual, but since the workingprinciple of these models is diffusion across a membrane, thisclassification seems meaningful.

    Artificial membranes

    In dialysis models with artificial membranes, different bacteriavessels were used, such as tubes,85 (square) vessels,86 separatingfunnels,87,88 artificial kidneys,8991 a plexiglass chamber withchangeable membrane filters,92,93 a hollow t-tube94,95 andhollow fibres.96 These models can have adjacent peripheral

    and central compartments or compartments embedded ineach other.

    In the adjacent setting, as in the model of Drugeon et al.,87

    the upper and lower part of a separating funnel are part of anentire loop (peripheral compartment), which also runs throughone part of the dialysis unit. The other part of the dialysis unitforms the central compartment; the dialysis unit enables theexchange of the drug and the medium. Continuous dilution ofthe central compartment decreases the drug concentration(Figure 2g).87,88 Toothaker et al.86 horizontally separated avessel by a haemodialysis membrane in two parts. One part con-tains fresh medium and the antibiotic, and the other partincludes the bacteria (Figure 2g).

    In theembedded setting, Guggenbichleret al.89 work with an

    artificial kidney, the inner part of which is connected to the bac-teria vessel (Figure 2h). Shah utilizes a plexiglass chamber withchangeable membrane filters at both ends as the bacteria com-partment ( peripheral compartment).93 This chamber is placed inan outer chamber with medium (central compartment;Figure 2h). The same model was adapted by Garrison et al.,94

    who modified the inner chamber to a hollow t-tube. In thehollow fibre model by Zinner et al.,96 the tubing of the centralcompartment includes a bundle of artificial capillaries. Thesecapillaries consist of polysulphone fibres, permeable for drugand medium, and are continuously flushed with medium. Thefibres pass through a vessel with bacteria (peripheral

    Dynamic models

    Dilution modelsDialysis/

    diffusion models

    Naturalbarrier

    Artificialbarrier

    Working principle

    Prevention of bacterial loss

    No YesYes

    Simpledilution

    Substitution

    Direct adding ordirect removingof medium

    Adding andremovingof medium

    Membrane material

    Figure 1. Detailed overview on dynamic in vitro models.

    Review

    192

  • 7/21/2019 bacterias anaerobias0f

    8/16

    compartment; Figure 2i). The hollow fibre model was furtherdeveloped by Blaser et al.,52,97 who added a second and morebacteria vessels.20,52,97 101 Al-Asadi et al.85 use two tubesclamped together, separated by a membrane, with drug in one(central compartment) and bacteria in the other tube (peripheralcompartment). After a finite time of drug diffusion from the drugto the bacteria tube, fresh medium is pumped into the bacteriatube and leads to a decrease of the drug concentration. In con-trary to all other dialysis models, here the bacteria compartmentitself is flushed (Figure 2j).

    Natural membranes

    The principle of dialysis models with natural membranes isalmost the same as for those with artificial membranes. Bacteria

    are captured behind a barrier and the drug has to pass thebarrier to reach the bacteria. Haller102 described a tissueculture model, where tissue cells are grown on a dialysis ultrafil-ter until a continuous layer is formed. The membrane, consistingof the filter and the tissue cell layer, is placed on a cylinder(central compartment). Another cylinder located above servesas the peripheral compartment with bacteria. The antibiotic isadministered in the lower part of the chamber by syringes anddiffuses through the cells to the upper part (Figure 2g). Themodel was suggested to investigate the penetration of thedrug through intercellular spaces and was later used by numer-ous groups to investigate drugbacteria effects.103107 An intra-cellular model implementing tissue cultures in PDin vitromodelsis presented by Hulten et al.103 Tissue cells are grown in insertsin a glass chamber, similar to a closed Petri dish (central

    Static Model (No. I)(a)

    B

    Dynamic dilution models (No. II)

    (b) stepwise substitution (1.)continuous simple dilution (2.)

    (c) continuous simple dilution,multi compartments

    BR W

    or1.

    2.BR W

    B

    Dynamic dilution models (No. IIIa)

    (d) stepwise simple dilution (1.)continuous dilution without outlet (2.)

    BR

    (e) stepwise substitution with filters, e.g.models by Nolting (1.+3.) or Haller

    (1.

    +

    4.)continuous dilution with filters, modelby Lowdin (2.+4.)

    (f) continuous dilution with filters,multi compartments

    BR W

    BR W

    or1.

    2.

    or1.

    2.or

    3.

    4.

    Figure 2. Schematic depiction of settings of in vitro models at the beginning of an experiment.

    Review

    193

    JA

  • 7/21/2019 bacterias anaerobias0f

    9/16

    compartment). The cells had previously been infected withintracellular-growing bacteria (peripheral compartment). Ametal rack for permeable cell culture inserts facilitates thetissue cell growth in the glass chamber. The cell membranesoperate as dialysis membranes and the cells are continuouslyflushed with fresh medium (Figure 2h). The drug has to passthe cell membrane to reach the bacteria. The bacteria can becounted after destruction of the cells.103109

    Experimental implementation ofin vivo routesof administration

    In vitromodels can be used to simulate different routes of drugadministration in patients. Generally, in dilution models the drugcan be added directly to the culture vessel or into an additionalvessel between the reservoir and the culture vessel, simulatingno (i.e. bolus administration) or first-order absorption (i.e. extra-vascular administration), respectively. From the additional vessel,the drug is transported with the medium into the culture vesseland into the waste. Zero-order absorption (i.e. infusion) of the

    drug can be achieved by adding the drug to the reservoir. Drug-containing medium is transported to the culture vessel and fromthere into the waste. The end of absorption in this case can berealized by exchange of the drug-containing reservoir to a drug-free reservoir (Figure 3a).

    Simulation of in vivo routes of drug administration in dialysismodels is the same as in dilution models. The drug is transportedwith the flowing medium to the central compartment (Figure 3b).From there it diffuses to the peripheral compartment. In all scen-arios the drug concentrations are suggested to follow in vivoabsorption/PK.21,57,62 Determinations of drug concentrations insamples from the culture vessel should support this assumption.

    Dynamic dialysis/diffusion models (No. IIIb)

    (g) Models with adjacent peripheral and centralcompartments with

    artificial membranes, models by e.g.Drugeon, Toothaker

    natural membranes, e.g. tissue culturemodel by Haller

    Models with embedded peripheralcompartments in central compartments with

    artificial membranes, e.g.Guggenbichler, Shah

    natural membranes, e.g. intracellularmodel by Hulten, fibrin clot modelby McGrath

    R WCentral

    B Peripheral

    BPeripheral

    R W

    Central

    (i) Hollow fibre model with artificial membrane Special case:Model by Al Asadi with artificial membrane

    R WCentral

    B Peripheral

    R WBPeripheral Central

    Caption:

    B culture vessel with bacteriaR reservoirW waste

    flow directionstepwise medium flowcontinuous medium flowfiltersemi-permeable membrane, i.e. permeable for drug and medium, not for bacteria

    (h)

    (j)

    Figure 2. Continued

    Table 5. Types of membranes in dialysis models (alphabetical order)

    Artificial membranes Natural membranes

    material ref. material ref.

    cellulose acetate 85 agarose gel 130

    haemodialysis membranes 86 cells 102

    polycarbonate 52, 94, 97 cell membranes 102 105

    polysulphone 96 fibrin 131

    regenerated cellulose 89 slime 47, 132

    synthetic regenerated cellulose

    ester

    95

    Review

    194

  • 7/21/2019 bacterias anaerobias0f

    10/16

    Applications

    A substantial number ofin vitroPD models have been developedto simulate specific conditions. Even if not all of these modelscan imitate the designated PK profiles, they are useful tools forspecific conditions. In Table 2, the models are grouped by their

    main aspects and may appear in different categories.

    Relevance and perspectives

    For the approval and rational use of antibiotics in pharmacother-apy, pre-clinical investigations will have to focus more on PK/PDinvestigations in the future. In this respect, in vitro modelsmight present a valuable predictive tool.1 A standardized meth-odology for use in pre-clinical research would provide a valuabletool for the optimization of dosing strategies.

    Generally, in vitromodels have several advantages comparedwithin vivoanimal studies: they are more flexible and adaptable

    to different conditions, and are less cost- and resource-intensive.Additionally, the relatively high inocula and volumes in in vitromodels allow better studies of resistance, because of the highermutation frequency than in animals.110 The PK properties of thedrug of interest can be applied in vitro and the time course ofan antimicrobial agent can be monitored exactly. On the

    other hand, in vitro models need special conditions, such asa temperature-controlled environment, and the risk of contami-nation of the culture vessel with external bacteria increases thelonger the experiment lasts.111 Since in vitro models cannotmimic all in vivo conditions,112 such as immunological factors(e.g. host defence mechanisms), the pathology of the infection,and the virulence and metabolic behaviour of a pathogen,1 thederived PD parameters cannot directly be transferred to thein vivosituation. Thein vivogrowth environment is different fromthe in vitro one. This may lead to phenotypic differencesbetween bacteria grownin vitroandin vivo.113 In general,in vitrobacterial growth is much faster than thatin vivo.38,114,115 Hence,

    (a) In dilution models with the example of a continuous simple dilution model

    BR W

    DD D

    A

    or

    0-order absorption

    Infusion

    0-order absorption

    Infusion

    Extravascularadministration

    1st-order absorption

    Extravascular

    administration1st-order absorption

    No absorption

    Bolus administration

    No absorption

    Bolus administration

    (b) In dialysis/diffusion models with the example of an embedded peripheralcompartment

    BPeripheral

    R W

    Central

    DD D

    A

    or

    D drugR reservoirA additional vessel, mimicking absorption (optional)B culture vessel with bacteriaW waste

    Figure 3. Schematic depiction of in vitro implementation of the differentin vivo routes of administration.

    Review

    195

    JA

  • 7/21/2019 bacterias anaerobias0f

    11/16

    a stronger competition for nutrients can lead to a higher pro-duction of antimicrobial drug targets, resulting in a higher suscep-tibility in vitro.43,114 In spite of this, in vitro models allow goodprediction of bacterial growthin vivoand comparison of differentdosingregimensofonedrugaswellascomparisonbetweendiffer-ent drugs. Finally, theycontribute to dose optimization.18,26,100,112

    The choice of a specific in vitro model is determined by theobjectives of a PK/PD study as well as the advantages and disad-vantages ofin vitromodels. Staticin vitromodels are extensivelyused as they are easy to handle and well investigated. Theyprovide basic information on the interaction between the anti-biotic and bacteria. In contrast to these favourable economicaspects is the unrealistic nature of the unchanged drug andmedium in static models,41 which fail to mirror two importantaspects of in vivo conditions, namely exchange of nutrientsand dilution of the drug. They are not useful for prolonged treat-ment studies, because nutrient depletion, space limitations andtoxic metabolites lead to growth restrictions.38 In our opinion,static models should be used as the starting point for PDstudies of the effect over time. The quick-and-easy findings

    from static models are useful preliminary knowledge fordynamic investigations.

    Dynamic in vitromodels represent the in vivoconditions withrespect to the changing drug and medium much moreclosely.14,41 Beside this, dynamic models also enable prolongedtreatment studies (up to 5,116,117 10118,119 and also 15 days120)with multiple dosing.52,71 Associated potential problems includethe Haag factor58 and membrane blockage. On the otherhand, they require large volumes of growth medium for changingthe drug concentration according to the half-lives of the drug.Hence, antibiotics with a long half-life have very low flow rates,a low volume of medium replacement and thus nutrientdepletion, and an increase of toxic metabolites. The two mainprinciples for changing drug concentrations have been developed

    and are both in use: dilution models and dialysis models.Dilution models can imitate virtually allin vivoPK profiles. The

    drug and bacteria are in one compartment, so the bacteria aredirectly exposed to the designated drug concentration. Hence,these models apply the designated PK drug profile to the bac-teria, but they should be monitored. The early dilution modelswere designed without bacterial retention as open models. Thebacterial loss always has to be corrected.49,50 In the Grassomodel,57 the bacteria leave the culture vessel with the outgoingmedium and toxic waste is diluted, but not considered. Addition-ally, bacterial aggregation and adherence to the vessel wall wasfound, where the bacterial populations should not be detect-able.58 In the model by Murakawa et al.40 the bacteria are dis-tributed into the second compartment and also eliminated into

    the waste. This complicates corrections for accurate bacterialconcentrations and, therefore, reliable predictions of the antibac-terial effect. The inserted bacteria filter causes new problems:the filter often becomes blocked with bacteria the longer theexperiment lasts.85,121 The implementation of pre-filters or stir-rers has provided potential solutions,55,82 but also open modelshave been reused. This has meant a step back regarding theloss of bacteria.121 In the later developed closed dilutionmodels, the bacterial backgrowth into the reservoir presentedanother problem. Dilution models with stepwise substitution(and filters), such as the syringe model by Nolting et al.,45 areeasily practicable, but need even more laborious effort than

    static models. They do not offer a continuous dilution and, there-with, not the same exposition profile for bacteria as in vivo.Nevertheless, it is possible to achieve more realistic resultsthan with static models.

    Dialysis models are extensively used, as well. Their mainadvantage is the closed system, whereby bacteria cannotescape and no further filters need to be installed. Dialysismodels enable simultaneous investigations of different bacterialstrains in separated vessels, but in one model.97 However, bac-teria accumulate at the membrane, which might becomeblocked (as in the case of dilution models).52,85 Furthermore,the changing drug concentration in the bacteria compartmentdoes not necessarily follow the designated PK profile,14 sincethe drug has to pass a barrier. Diffusion of a drug is a first-orderprocess. The extent of drug transfer across the barrier dependson the site of membrane permeation and varies with time.This means there is a specific concentration gradient betweenthe drug concentration in the central compartment and periph-eral compartment at each timepoint. Unfortunately, only a fewgroups have determined the drug concentration in the bacteria

    compartment,52,53,85,86,89,92,94,97 99 where the concentrationgradient was confirmed. The gradient can be improved byhigher circulation (addition of pumps) and contraflow of themedium in the central compartment and peripheral compart-ment.52,53,97 The early dialysis models with artificial membranesas well as most with natural membranes suffered from a smallsurface area of the membrane to volume ratio. This led to dimin-ished diffusion or membrane blockage.85,87,93 Later, the exactratio between the membrane surface area and the volume ofthe peripheral compartment was estimated and changed, e.g.by Vance-Bryanet al.,92,122 for the Shah model.93 In intracellularmodels, determination of the drug concentration is even moreproblematic as the site of action is inside the cells. Here, theflow is not directed for an optimal exchange between extra-

    and intracellular fluid, which may lead to different PK profilesthat the bacteria will be exposed to. Only with special equipmentand procedures such as fluorescence microscopy is it possible tomeasure the drug concentration inside the cells. So, the effect isoften related to the drug decline in the central compartment,because of easier determination.

    In summary, in spite of their simplicity, static models will stillplay an important role for antibacterial PK/PD studies in thefuture, but should be regarded as a starting point. For morecomplex PK designs, dynamic models will be more importantand their use will hopefully increase. The dilution models,such as the Grasso model,57 have existed since the 1970sand have been intensively diversified. Almost at the sametime, dialysis models have been introduced and, later,

    improved. Currently, the ratio of using dialysis or dilutionmodels is balanced. Both types of dynamic models have beenfurther developed in the past and are presented in thecurrent literature. New developments combine the ideas of aone-compartment dilution model with filters and a two-compartment dialysis model, resulting in a computer-controlledsemi-automated in vitro model for industrial purposes.111 Infuture, this trend of combining models for different purposes,as well as automation, might lead to more frequent use and,eventually, they might become an inherent part of drug discov-ery and development. Comprehensive understanding of the PDof antibiotics should facilitate the development of rational

    Review

    196

  • 7/21/2019 bacterias anaerobias0f

    12/16

    dosing schedules for patients, resulting in improved therapy andlower mortality. We hope that in the future in vitro models willincreasingly be used to define the PK/PD characteristics of anti-biotics and will serve to complement the data from clinicaltrials.

    FundingThis work was partially supported by a grant from the Dr. August undDr. Anni Lesmueller-Stiftung, Germany.

    Transparency declarationsThe authors do not have any financial, commercial or proprietary interestin any drug, device or equipment mentioned in this paper.

    References1 EMEA.Points to Consider on Pharmacokinetics and Pharmacodynamics in

    the Development of Antibacterial Medicinal Products. CPMP/EWP/2655/99.2000. http://www.emea.europa.eu/pdfs/human/ewp/265599en.pdf (3December 2009, date last accessed).

    2 Holford NH, Sheiner LB. Kinetics of pharmacologic response.PharmacolTher1982; 16: 14366.

    3 FDA. Guidance for Industry. Developing Antimicrobial DrugsGeneralConsiderations for Clinical Trials (Draft Guidance). 1998. http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm070983.pdf (19 May 2009, date last accessed).

    4 Fantin B, Carbon C. In vivo antibiotic synergism: contribution of animalmodels.Antimicrob Agents Chemother1992; 36: 90712.

    5 Dudley MN, Griffith D. Animal models of infection for the study ofantibiotic pharmacodynamics. In: Nightingale CH, Murakawa T,Ambrose PG, eds. Antimicrobial Pharmacodynamics in Theory and

    Clinical Practice. New York, NY: Marcel Dekker, Inc., 2002; 6798.6 Craig WA. Pharmacokinetic/pharmacodynamic parameters: rationalefor antibacterial dosing of mice and men. Clin Infect Dis1998;26: 1 10.

    7 Hickey E. Tools to define the relevance of PK/PD parameters to theefficacy, toxicity and emergence of resistance of antimicrobials. CurrOpin Drug Discov Devel 2007; 10: 4952.

    8 Drusano GL, Louie A, Deziel Met al. The crisis of resistance: identifyingdrug exposures to suppress amplification of resistant mutantsubpopulations.Clin Infect Dis 2006; 42: 52532.

    9 Mouton JW, Vinks AA. Pharmacokinetic/pharmacodynamic modellingof antibacterials in vitro and in vivo using bacterial growth and killkinetics: the minimum inhibitory concentration versus stationaryconcentration.Clin Pharmacokinet2005; 44: 20110.

    10 Mouton JW, Dudley MN, Cars O et al. Standardization of

    pharmacokinetic/pharmacodynamic (PK/PD) terminology foranti-infective drugs. Int J Antimicrob Agents 2002; 19: 3558.

    11 Mouton JW, Dudley MN, Cars O et al. Standardization ofpharmacokinetic/pharmacodynamic (PK/PD) terminology foranti-infective drugs: an update.J Antimicrob Chemother2005;55: 6017.

    12 Barger A, Fuhst C, Wiedemann B. Pharmacological indices in antibiotictherapy. J Antimicrob Chemother2003; 52: 8938.

    13 Blondeau JM, Hansen G, Metzler Ket al. The role of PK/PD parametersto avoid selection and increase of resistance: mutant preventionconcentration.J Chemother2004; 16 Suppl 3: 119.

    14 Grasso S. Historical review of in-vitro models.J Antimicrob Chemother1985;15 Suppl A: 99102.

    15 Firsov AA, Zinner SH, Lubenko IY. In vitro dynamic models as tools topredict antibiotic pharmacodynamics. In: Nightingale CH, Ambrose PG,Drusano GL et al., eds. Antimicrobial Pharmacodynamics in Theory andClinical Practice, 2nd edn. New York, NY: Informa Healthcare, 2007;4578.

    16 Rybak MJ, Allen GP, Hershberger E. In vitro antibiotic

    pharmacodynamic models. In: Nightingale CH, Murakawa T, AmbrosePG, eds. Antimicrobial Pharmacodynamics in Theory and Clinical Practice.New York, NY: Marcel Dekker, Inc., 2002; 4166.

    17 Li RC, Zhu ZY. In vitro models for prediction of antimicrobial activity: apharmacokinetic and pharmacodynamic perspective. J Chemother1997;9Suppl 1: 5563.

    18 MacGowan A, Bowker K. Developments in PK/PD: optimising efficacyand prevention of resistance. A critical review of PK/PD in in vitromodels.Int J Antimicrob Agents 2002; 19: 2918.

    19 MacGowan A, Rogers C, Bowker K. In vitro models, in vivo models, andpharmacokinetics: what can we learn from in vitro models? Clin Infect Dis2001;33 Suppl 3: 21420.

    20 Blaser J. In-vitro model for simultaneous simulation of the serumkinetics of two drugs with different half-lives. J Antimicrob Chemother

    1985;15 Suppl A: 12530.21 Rowe EL, Morozowich W. A simple dilution analog computer forsimulation of drug distribution processes. J Pharm Sci 1969;58: 13758.

    22 Derendorf H, Meibohm B. Modeling of pharmacokinetic/pharmacodynamic (PK/PD) relationships: concepts and perspectives.Pharm Res 1999; 16: 17685.

    23 Czock D, Keller F. Mechanism-based pharmacokineticpharmacodynamic modeling of antimicrobial drug effects.J Pharmacokinet Pharmacodyn 2007; 34: 72751.

    24 Schuck EL, Derendorf H. Pharmacokinetic/pharmacodynamicevaluation of anti-infective agents. Expert Rev Anti Infect Ther2005; 3:36173.

    25 Meibohm B, Derendorf H. Basic concepts of pharmacokinetic/pharmacodynamic (PK/PD) modelling. Int J Clin Pharmacol Ther1997;

    35: 40113.26 Frimodt-Moller N. How predictive is PK/PD for antibacterial agents?IntJ Antimicrob Agents 2002; 19: 3339.

    27 Yano Y, Oguma T, Nagata Het al. Application of logistic growth modelto pharmacodynamic analysis of in vitro bactericidal kinetics. J Pharm Sci1998;87: 117783.

    28 Nielsen EI, Viberg A, Lowdin E et al. Semimechanisticpharmacokinetic/pharmacodynamic model for assessment of activityof antibacterial agents from time kill curve experiments. AntimicrobAgents Chemother2007; 51: 12836.

    29 Meagher AK, Forrest A, Dalhoff A et al. Novel pharmacokineticpharmacodynamic model for prediction of outcomes with anextended-release formulation of ciprofloxacin. Antimicrob AgentsChemother2004; 48: 20618.

    30 Nikolaou M, Schilling AN, Vo G et al. Modeling of microbial populationresponses to time-periodic concentrations of antimicrobial agents. AnnBiomed Eng 2007; 35: 145870.

    31 Nikolaou M, Tam VH. A new modeling approach to the effect ofantimicrobial agents on heterogeneous microbial populations. J MathBiol2006;52: 15482.

    32 Zhi J, Nightingale CH, Quintiliani R. A pharmacodynamic model for theactivity of antibiotics against microorganisms under nonsaturableconditions.J Pharm Sci 1986; 75: 10637.

    33 Firsov AA, Vostrov SN, Shevchenko AA et al. Parameters of bacterialkilling and regrowth kinetics and antimicrobial effect examined interms of area under the concentrationtime curve relationships: action

    Review

    197

    JA

  • 7/21/2019 bacterias anaerobias0f

    13/16

    of ciprofloxacin against Escherichia coli in an in vitro dynamic model.Antimicrob Agents Chemother1997; 41: 12817.

    34 Firsov AA, Vostrov SN, Shevchenko AAet al. A new approach to in vitrocomparisons of antibiotics in dynamic models: equivalent area under thecurve/MIC breakpoints and equiefficient doses of trovafloxacin andciprofloxacin against bacteria of similar susceptibilities. Antimicrob

    Agents Chemother1998;42: 28417.

    35 Campion JJ, McNamara PJ, Evans ME. Pharmacodynamic modeling ofciprofloxacin resistance in Staphylococcus aureus. Antimicrob AgentsChemother2005;49: 20919.

    36 Campion JJ, Chung P, McNamara PJ et al. Pharmacodynamicmodeling of the evolution of levofloxacin resistance in Staphylococcusaureus. Antimicrob Agents Chemother2005; 49: 218999.

    37 Chung P, McNamara PJ, Campion JJ e t al. Mechanism-basedpharmacodynamic models of fluoroquinolone resistance inStaphylococcus aureus. Antimicrob Agents Chemother 2006; 50:295765.

    38 Gilbert P. The theory and relevance of continuous culture.J AntimicrobChemother1985;15 Suppl A: 16.

    39 Bernaerts K, Dens E, Vereecken K et al. Concepts and tools forpredictive modeling of microbial dynamics. J Food Prot 2004; 67:204152.

    40 Murakawa T, Sakamoto H, Hirose T et al. New in vitro kinetic model forevaluating bactericidal efficacy of antibiotics. Antimicrob AgentsChemother1980;18: 37781.

    41 Mueller M, de la Pena A, Derendorf H. Issues in pharmacokinetics andpharmacodynamics of anti-infective agents: kill curves versus MIC.Antimicrob Agents Chemother2004; 48: 36977.

    42 Derendorf H, Hochhaus G. Handbook of Pharmacokinetic/Pharmacodynamic Correlation. Boca Raton, FL: CRC Press Inc., 1995.

    43 Garrett ER, Miller GH, Brown MR. Kinetics and mechanisms of action ofantibiotics on microorganisms. V. Chloramphenicol and tetracyclineaffected Escherichia coli generation rates. J Pharm Sci 1966; 55:

    593600.44 Treyaprasert W, Schmidt S, Rand KH et al. Pharmacokinetic/pharmacodynamic modeling of in vitro activity of azithromycin againstfour different bacterial strains.Int J Antimicrob Agents2007;29: 26370.

    45 Nolting A, Dalla Costa T, Rand KH et al. Pharmacokineticpharmacodynamic modeling of the antibiotic effect of piperacillin invitro.Pharm Res 1996; 13: 916.

    46 Scaglione F, Demartini G, Dugnani S et al. A new model examiningintracellular and extracellular activity of amoxicillin, azithromycin, andclarithromycin in infected cells. Chemotherapy1993; 39: 41623.

    47 Darouiche RO, Dhir A, Miller AJ et al. Vancomycin penetration intobiofilm covering infected prostheses and effect on bacteria. J Infect Dis1994;170: 7203.

    48 Garrett ER, Wright OK, Miller GHet al. Quantification and prediction of

    the biological activities of chloramphenicol analogs by microbial kinetics.J Med Chem1966; 9: 2038.

    49 Keil S, Wiedemann B. Mathematical corrections for bacterial loss inpharmacodynamic in vitro dilution models. Antimicrob AgentsChemother1995;39: 10548.

    50 White CA, Toothaker RD, Smith ALet al. Correction for bacterial loss inin vitro dilution models. Antimicrob Agents Chemother 1987; 31:185960.

    51 den Hollander JG, Mouton JW, Verbrugh HA. Use of pharmacodynamic parameters to predict efficacy of combinationtherapy by using fractional inhibitory concentration kinetics. AntimicrobAgents Chemother1998;42: 7448.

    52 Blaser J, Stone BB, Zinner SH. Two compartment kinetic model withmultiple artificial capillary units. J Antimicrob Chemother 1985; 15Suppl A: 1317.

    53 Mouton JW, den Hollander JG. Killing of Pseudomonas aeruginosaduring continuous and intermittent infusion of ceftazidime in an invitro pharmacokinetic model. Antimicrob Agents Chemother 1994; 38:

    9316.

    54 Garrett ER, Nolte H. Kinetics and mechanisms of drug action onmicroorganisms. XIV. The action of fluorouracil, other uracils andderived nucleosides on the microbial kinetics of Escherichia coli.Chemotherapy1972;17: 81108.

    55 Lowdin E, Odenholt I, Bengtsson S et al. Pharmacodynamic effects ofsub-MICs of benzylpenicillin against Streptococcus pyogenes in a newlydeveloped in vitro kinetic model. Antimicrob Agents Chemother 1996;40: 247882.

    56 Nishida M, Murakawa T, Kamimura Tet al. Laboratory evaluation ofFR10612, a new oral cephalosporin derivative. J Antibiot (Tokyo) 1976;29: 44459.

    57 Grasso S, Meinardi G, de Carneri I et al. New in vitro model to studythe effect of antibiotic concentration and rate of elimination onantibacterial activity. Antimicrob Agents Chemother1978; 13: 5706.

    58 Haag R, Lexa P, Werkhauser I. Artifacts in dilution pharmacokineticmodels caused by adherent bacteria. Antimicrob Agents Chemother1986;29: 7658.

    59 Gerber AU, Wiprachtiger P, Stettler-Spichiger U et al. Constantinfusions vs. intermittent doses of gentamicin against Pseudomonasaeruginosa in vitro. J Infect Dis 1982; 145: 55460.

    60 Satta G, Cornaglia G, Foddis G et al. Evaluation of ceftriaxone andother antibiotics against Escherichia coli, Pseudomonas aeruginosa, andStreptococcus pneumoniae under in vitro conditions simulating those ofserious infections. Antimicrob Agents Chemother1988; 32: 55260.

    61 Firsov AA, Chernykh VM, Kuznetsova SM et al. A dynamic system forthe in vitro study of the kinetics of the antimicrobial effect ofantibiotics in pharmacokinetic changes in their concentration. AntibiotMed Biotekhnol1985; 30: 3643.

    62 Firsov AA, Nazarov AD, Chernykh VM et al. Validation of optimalampicillin/sulbactam ratio in dosage forms using in-vitro dynamicmodel.Drug Dev Ind Pharm 1988; 14: 242542.

    63 MacGowan AP, Bowker KE, Noel AR. Pharmacodynamics of theantibacterial effect and emergence of resistance to tomopenem,formerly RO4908463/CS-023, in an in vitro pharmacokinetic model ofStaphylococcus aureus infection. Antimicrob Agents Chemother 2008;

    52: 14016.

    64 MacGowan AP, Bowker KE, Wootton M et al. Activity of moxifloxacin,administered once a day, againstStreptococcus pneumoniaein an in vitropharmacodynamic model of infection. Antimicrob Agents Chemother1999;43: 15604.

    65 Haller I. Combined action of decreasing concentrations of azlocillinand sisomicin on Pseudomonas aeruginosa as assessed in a dynamic invitro model. Infection1982; 10 Suppl 3: S22933.

    66 MacGowan AP, Rogers CA, Holt HA et al. Activities of moxifloxacinagainst, and emergence of resistance in, Streptococcus pneumoniaeand Pseudomonas aeruginosa in an in vitro pharmacokinetic model.Antimicrob Agents Chemother2003; 47: 108895.

    67 Nies BA. Comparative activity of cefixime and cefaclor in an in vitromodel simulating human pharmacokinetics. Eur J Clin Microbiol InfectDis1989; 8: 55861.

    68 Allen GP, Kaatz GW, Rybak MJ. Activities of mutant preventionconcentration-targeted moxifloxacin and levofloxacin against

    Review

    198

  • 7/21/2019 bacterias anaerobias0f

    14/16

  • 7/21/2019 bacterias anaerobias0f

    15/16

    103 Hulten K, Rigo R, Gustafsson I et al. New pharmacokinetic in vitromodel for studies of antibiotic activity against intracellularmicroorganisms.Antimicrob Agents Chemother1996; 40: 272731.

    104 Birkness KA, Swisher BL, White EH et al. A tissue culture bilayermodel to study the passage of Neisseria meningitidis. Infect Immun1995;63: 4029.

    105 Birkness KA, Deslauriers M, Bartlett JHet al. An in vitro tissue culturebilayer model to examine early events in Mycobacterium tuberculosisinfection.Infect Immun1999; 67: 6538.

    106 Shaw JH, Hayes F, Brooks GFet al. Development of a tissue culturemodel for gonococcal invasion. Antonie Van Leeuwenhoek 1987; 53:48591.

    107 Shaw JH, Falkow S. Model for invasion of human tissue culture cellsbyNeisseria gonorrhoeae. Infect Immun1988; 56: 162532.

    108 Gaillard JL, Berche P, Mounier J et al. In vitro model of penetrationand intracellular growth of Listeria monocytogenes in the humanenterocyte-like cell line Caco-2. Infect Immun1987; 55: 28229.

    109 Fattorini L, Li B, Piersimoni C et al. In vitro and ex vivo activities ofantimicrobial agents used in combination with clarithromycin, with or

    without amikacin, against Mycobacterium avium. Antimicrob AgentsChemother1995;39: 6805.

    110 Craig WA, Andes DR. In vivo pharmacodynamics of ceftobiproleagainst multiple bacterial pathogens in murine thigh and lung infectionmodels.Antimicrob Agents Chemother2008; 52: 34926.

    111 Wang L, Wismer MK, Racine F et al. Development of an integratedsemi-automated system for in vitro pharmacodynamic modelling.J Antimicrob Chemother2008; 62: 10707.

    112 Lorian V. In vitro simulation of in vivo conditions: physical state ofthe culture medium. J Clin Microbiol 1989;27: 24036.

    113 Dalhoff A. Differences between bacteria grown in vitro and in vivo.J Antimicrob Chemother1985; 15 Suppl A: 17595.

    114 Brown MR, Collier PJ, Gilbert P. Influence of growth rate onsusceptibility to antimicrobial agents: modification of the cell envelope

    and batch and continuous culture studies. Antimicrob AgentsChemother1990;34: 16238.

    115 Lorian V. Differences between in vitro and in vivo studies.AntimicrobAgents Chemother1988;32: 16001.

    116 Tam VH, Kabbara S, Vo G et al. Comparative pharmacodynamics ofgentamicin against Staphylococcus aureus and Pseudomonasaeruginosa. Antimicrob Agents Chemother2006; 50: 262631.

    117 Tam VH, Schilling AN, Neshat S et al. Optimization of meropenemminimum concentration/MIC ratio to suppress in vitro resistance ofPseudomonas aeruginosa.Antimicrob Agents Chemother2005; 49: 49207.

    118 Gumbo T, Louie A, Deziel MR et al. Selection of a moxifloxacin dosethat suppresses drug resistance in Mycobacterium tuberculosis, by use ofan in vitro pharmacodynamic infection model and mathematicalmodeling.J Infect Dis 2004;190: 164251.

    119 Louie A, Brown DL, Liu W et al. In vitro infection model characterizingthe effect of efflux pump inhibition on prevention of resistance tolevofloxacin and ciprofloxacin in Streptococcus pneumoniae. AntimicrobAgents Chemother2007;51: 39884000.

    120 Louie A, Heine HS, Kim K et al. Use of an in vitro pharmacodynamicmodel to derive a linezolid regimen that optimizes bacterial kill andprevents emergence of resistance in Bacillus anthracis. AntimicrobAgents Chemother2008;52: 248696.

    121 Venisse N, Gregoire N, Marliat M et al. Mechanism-basedpharmacokineticpharmacodynamic models of in vitro fungistatic andfungicidal effects against Candida albicans. Antimicrob AgentsChemother2008;52: 93743.

    122 Vance-Bryan K, Larson TA, Rotschafer JC et al. Investigation of theearly killing of Staphylococcus aureus by daptomycin by using an invitro pharmacodynamic model. Antimicrob Agents Chemother1992; 36:23347.

    123 Eng RH, Smith SM, Cherubin CEet al. Evaluation of two methods forovercoming the antibiotic carry-over effect. Eur J Clin Microbiol Infect Dis

    1991;10: 348.

    124 den HollanderJG, Mouton JW,Bakker-WoudenbergIA etal. Enzymaticmethod for inactivation of aminoglycosides during measurement ofpostantibiotic effect.Antimicrob Agents Chemother1996;40: 48890.

    125 den Hollander JG, Mouton JW, van Goor MP et al. Alteration ofpostantibiotic effect during one dosing interval of tobramycin,simulated in an in vitro pharmacokinetic model. Antimicrob AgentsChemother1996; 40: 7846.

    126 Hanberger H, Svensson E, Nilsson M et al. Effects of imipenem onEscherichia coli studied using bioluminescence, viable counting andmicroscopy.J Antimicrob Chemother1993;31: 24560.

    127 Fang W. Quantification ofStaphylococcus aureusandEscherichia coliin the liquid medium by fluorimetry and its use in phagocytosis assay.J Appl Bacteriol1996; 80: 57782.

    128 Fey A, Eichler S, Flavier S et al. Establishment of a real-timePCR-based approach for accurate quantification of bacterial RNAtargets in water, using Salmonella as a model organism. Appl EnvironMicrobiol2004; 70: 3618 23.

    129 Smith CJ, Osborn AM. Advantages and limitations of quantitativePCR (Q-PCR)-based approaches in microbial ecology. FEMS Microbiol Ecol2009;67: 620.

    130 Tomita T, Ohara-Nemoto Y, Moriyama H et al. A novel in vitropharmacokinetic/pharmacodynamic model based on two-compartmentopen model used to simulate serum drug concentrationtime profiles.Microbiol Immunol 2007; 51: 56775.

    131 McGrath BJ, Kang SL, Kaatz GW et al. Bactericidal activities ofteicoplanin, vancomycin, and gentamicin alone and in combinationagainst Staphylococcus aureus in an in vitro pharmacodynamic

    model of endocarditis.Antimicrob Agents Chemother1994;38: 203440.132 Vergeres P, Blaser J. Amikacin, ceftazidime, and flucloxacillin againstsuspended and adherent Pseudomonas aeruginosa and Staphylococcusepidermidis in an in vitro model of infection. J Infect Dis 1992; 165:2819.

    133 Randolph JA, Buck RE, Price KE et al. Comparative bactericidal effectof ceforanide (BL-S 786) and five other cephalosporins in an in vitropharmacokinetic model. J Antibiot (Tokyo) 1979; 32: 72733.

    134 Mah TF, OToole GA. Mechanisms of biofilm resistance toantimicrobial agents.Trends Microbiol 2001; 9: 349.

    135 Nickel JC, Wright JB, Ruseska I et al. Antibiotic resistance ofPseudomonas aeruginosa colonizing a urinary catheter in vitro. Eur JClin Microbiol1985; 4: 2138.

    136 Prosser BL, Taylor D, Dix BA et al. Method of evaluating effects of

    antibiotics on bacterial biofilm. Antimicrob Agents Chemother1987; 31:15026.

    137 Nickel JC, Ruseska I, Wright JB et al. Tobramycin resistance ofPseudomonas aeruginosa cells growing as a biofilm on urinary cathetermaterial.Antimicrob Agents Chemother1985;27: 61924.

    138 Ellen RP, Lepine G, Nghiem PM. In vitro models that supportadhesion specificity in biofilms of oral bacteria.Adv Dent Res 1997; 11:3342.

    139 Leunisse C, van Weissenbruch R, Busscher HJ et al. The artificialthroat: a new method for standardization of in vitro experiments withtracheo-oesophageal voice prostheses. Acta Otolaryngol 1999; 119:6048.

    Review

    200

  • 7/21/2019 bacterias anaerobias0f

    16/16

    140 Kutlin A, Roblin PM, Hammerschlag MR. In vitro activities ofazithromycin and ofloxacin against Chlamydia pneumoniae in acontinuous-infection model. Antimicrob Agents Chemother 1999; 43:226872.

    141 Orme IM, Roberts AD, Furney SK et al. Animal and cell-culturemodels for the study of mycobacterial infections and treatment. Eur

    J Clin Microbiol Infect Dis 1994; 13: 9949.

    142 OGrady F, Mackintosh IP, Greenwood Det al. Treatment of "bacterialcystitis" in fully automatic mechanical models simulating conditions ofbacterial growth in the urinary bladder. Br J Exp Pathol1973;54: 28390.

    143 Greenwood D, OGrady F. An in vitro model of the urinary bladder.J Antimicrob Chemother1978; 4: 11320.

    144 Sano M, Kumamoto Y, Nishimura M et al. Inhibition of biofilmformation by clarithromycin (CAM) in an experimental model ofcomplicated bladder infectionin vitro study using automatedsimulation of urinary antimicrobial concentration. KansenshogakuZasshi 1994; 68: 130617.

    145 Greenwood D. An in-vitro model simulating the hydrokinetic aspectsof the treatment of bacterial cystitis. J Antimicrob Chemother1985; 15Suppl A: 1039.

    146 Eden T. Long-standing otitis media with effusiona convenientmodel for the study of antibiotic penetration to respiratory tractsecretions.Scand J Infect Dis Suppl 1985; 44: 4651.

    147 Palmer SM, Rybak MJ. An evaluation of the bactericidal activity ofampicillin/sulbactam, piperacillin/tazobactam, imipenem or nafcillinalone and in combination with vancomycin against methicillin-resistantStaphylococcus aureus (MRSA) in timekill curves with infected fibrinclots.J Antimicrob Chemother1997;39: 5158.

    148 Hershberger E, Rybak MJ. Activities of trovafloxacin, gatifloxacin,clinafloxacin, sparfloxacin, levofloxacin, and ciprofloxacin againstpenicillin-resistant Streptococcus pneumoniae in an in vitro infectionmodel.Antimicrob Agents Chemother2000; 44: 598 601.

    149 Sissons CH. Artificial dental plaque biofilm model systems. Adv DentRes1997; 11: 11026.

    150 Herles S, Olsen S, Afflitto J et al. Chemostat flow cell system: an invitro model for the evaluation of antiplaque agents. J Dent Res 1994;73: 174855.

    151 McDermid AS, McKee AS, Marsh PD. A mixed-culture chemostatsystem to predict the effect of anti-microbial agents on the oral flora:preliminary studies using chlorhexidine. J Dent Res 1987; 66: 131520.

    152 Herruzo-Cabrera R, Vizcaino-Alcaide MJ, Mayer RF et al. A new in vitromodel to test the effectiveness of topical antimicrobial agents. Use of anartificial eschar.Burns1992; 18: 358.

    153 Garrett ER, Brown MR. The action of tetracycline andchloramphenicol alone and in admixture on the growth ofEscherichiacoli.J Pharm Pharmacol 1963; 15 Suppl: 18591.

    154 Liu Q, Rand K, Derendorf H. Impact of tazobactam pharmacokineticson the antimicrobial effect of piperacillintazobactam combinations. IntJ Antimicrob Agents 2004; 23: 4947.

    155 LaPlante KL, Leonard SN, Andes DR et al. Activities of clindamycin,daptomycin, doxycycline, linezolid, trimethoprim sulfamethoxazole,and vancomycin against community-associated methicillin-resistant

    Staphylococcus aureus with inducible clindamycin resistance in murinethigh infection and in vitro pharmacodynamic models. AntimicrobAgents Chemother2008; 52: 215662.

    156 Allen GP, Cha R, Rybak MJ. In vitro activities of quinupristindalfopristin and cefepime, alone and in combination with variousantimicrobials, against multidrug-resistant staphylococci andenterococci in an in vitro pharmacodynamic model. Antimicrob AgentsChemother2002; 46: 260612.

    157 Akins RL, Rybak MJ. In vitro activities of daptomycin, arbekacin,vancomycin, and gentamicin alone and/or in combination againstglycopeptide intermediate-resistant Staphylococcus aureus in aninfection model. Antimicrob Agents Chemother2000; 44: 19259.

    158 Lim TP, Ledesma KR, Chang KT et al. Quantitative assessment ofcombination antimicrobial therapy against multidrug-resistant

    Acinetobacter baumannii. Antimicrob Agents Chemother 2008; 52:2898904.

    159 Tam VH, Schilling AN, Lewis RE et al. Novel approach tocharacterization of combined pharmacodynamic effects ofantimicrobial agents.Antimicrob Agents Chemother2004; 48: 431521.

    160 den Hollander JG, Horrevorts AM, van Goor ML et al. Synergismbetween tobramycin and ceftazidime against a resistant Pseudomonasaeruginosa strain, tested in an in vitro pharmacokinetic model.Antimicrob Agents Chemother1997; 41: 95100.

    161 Barchiesi F, Spreghini E, Tomassetti S e t al. Caspofungin incombination with amphotericin B against Candida parapsilosis.Antimicrob Agents Chemother2007; 51: 9415.

    162 Lewis RE, Wiederhold NP, Prince RAet al. In vitro pharmacodynamicsof rapid versus continuous infusion of amphotericin B deoxycholate

    against Candida species in the presence of human serum albumin.J Antimicrob Chemother2006;57: 28893.

    163 Lignell A, Johansson A, Lowdin Eet al. A new in-vitro kinetic modelto study the pharmacodynamics of antifungal agents: inhibition of thefungicidal activity of amphotericin B against Candida albicans byvoriconazole.Clin Microbiol Infect 2007; 13: 6139.

    164 Zabinski RA, Walker KJ, Larsson AJ et al. Effect of aerobic andanaerobic environments on antistaphylococcal activities of fivefluoroquinolones.Antimicrob Agents Chemother1995;39: 50712.

    165 Lewis RE, Kontoyiannis DP, Darouiche RO et al. Antifungal activity ofamphotericin B, fluconazole, and voriconazole in an in vitro model ofCandida catheter-related bloodstream infection. Antimicrob AgentsChemother2002; 46: 3499505.

    Review JA