relevance of the alexander project

Journal of Antimicrobial Chemotherapy (1996) 38, Suppl. A, 141-154

Relevance of the Alexander Project: phannacodynamic considerations

G. L. Drusano' and F. W. Goldstein*

"Division of Clinical Pharmacology, Albany Medical College, 47 New Scotland Avenue,Albany, NY 12208, USA; bLaboratoire de Microbiologie Medicale, Hopital Saint-Joseph,

185 rue Raymond Losserand, 75674 Paris Cedex 14, France

Application of pharmacodynamic principles for interpretation of data generated bythe Alexander Project is possible for /Mactam, quinolone and macrolide antibiotics.For /Mactams, the time that serum concentrations remain above the MIC of thepathogen (T > MIC) is the parameter most closely linked with outcome. It has beenshown that T > MIC need be only 50-60% of a dose interval. Since the MIC hasthe greatest influence on this parameter, a conservative estimate of activity woulduse the MICV The only /Mactam antibiotics in the Alexander Project for whichT > MIC» for the four major pathogens (Streptococcus pneumoniae, Haemophilusinfluenzae, Moraxella catarrhalis and Staphylococcus aureus) exceeded 50% of thedose interval were amoxycillin/clavulanate (500/125 mg) and ceftriaxone. Formacrolides, T > MIC is relevant for erythromycin and clarithromycin, but notazithromycin, for which AUC is the parameter most closely linked to outcome.Erythromycin, clarithromycin and azithromycin showed efficacy against M.catarrhalis only at MICV Quinolones (ciprofloxacin and ofloxacin), for which AUCis also the relevant phannacodynamic parameter, had the greatest activity against H.influenzae and M. catarrhalis at MIG», but were less effective against S. pneumoniaeand 5. aureus. Susceptibility data such as those provided by the Alexander Projectcan aid clinicians in choosing appropriate treatment for LRTI based onpharmacodynamic principles.

Introduction

A considerable body of work has emerged in the last 15 years linking different measuresof pharmacological exposure to an antimicrobial agent to the outcome observed in thesetting of a serious infection. These data have been developed in vivo in animal modelsystems, and in patients, and allow rational extrapolation from in-vitro data to expectedclinical utility.

There are two factors which influence potential utility of a drug in a specific clinicalsituation. The first is some measure of potency of the antibiotic for the pathogen inquestion (MIC, MBC, etc.). The second is whichever relationship between theconcentration-time profile and potency of the antibiotic is linked most robustly toclinical outcome (Time above MIC or MBC [T > MIC or T > MBC]; Peak/MIC orMBC; AUC/MIC or AUC/MBC).

The second factor has been documented extensively for /?-lactam drugs. Clearly, onthe basis of in-vitro data, animal model system data and, most importantly, clinical trialdata, the time the free antibiotic concentration remains above the MIC (MBC) is most

1410305-7453/96/38A14I + 14 $12.00/0 © 1996 The British Society for Antimicrobial Chemotherapy

Dow

nloaded from https://academ

ic.oup.com/jac/article/38/suppl_A/141/728098 by guest on 01 August 2022

142 G. L. Orusano and F. W. Goldstein

clearly linked to outcome (Drusano, 1988). /?-Lactam kill-rates are relativelyconcentration independent. That is, in the in-vitro situation, increasing theconcentration of the /Mactam has little effect on the kill rate observed. Indeed, for moststrains, there is only a mild increase in kill rate as the concentration increases from theMIC (or MBC) to four times the MIC (or MBC). After this, there is little furtherincrease in the rate of bacterial killing with increasing concentration.

These data employing /?-lactam agents are in stark contrast to data developedwith other classes of anti-infective drugs. Drugs such as aminoglycosides andfluoroquinolones have a highly concentration-dependent kill rate. Increasing exposureconcentrations results in major increases in kill rate over a wide range. For these agents,the total number of organisms killed is merely a path integral of concentration-dependent kill rate times. It is straightforward, then, that the area under theconcentration-time curve (AUC), itself an integral of the concentration over time, is thevariable most directly linked to organism kill and ultimate outcome for agents whichare concentration-dependent in kill rate.

What other data are there that support the suggestion that time above MIC is themost important pharmacological measure linked to successful therapy? Zinner et al.(1988) studied cefoperazone in an in-vitro hollow fibre system and these data clearlyindicated that the time that the drug concentration exceeded the MIC for the pathogenwas the measure most closely linked to outcome.

Convincing data indicating that T > MIC was important were generated by Gerberet al. (1983) in an animal model. In this system, mice were rendered neutropenic withcyclophosphamide, allowing the underlying relationship between drug exposure andresponse to be studied more easily. In the first of a series of publications from this group,ticarcillin was the antibiotic employed. Animals were infected with Pseudomonasaeruginosa in the thigh, and ticarcillin was given as a large dose every 3 h and comparedwith one-third of the same dose administered every hour. The difference in outcome isdisplayed in Figure 1. Clearly, the smaller dose, given more frequently so maximizingthe time that plasma concentrations exceed the MIC for the P. aeruginosa, wassignificantly more effective in controlling the infection at the primary site.

Later, this group examined another way of discerning the pharmacological variablemost closely linked to outcome (Leggett et al., 1989). Small numbers of animals wereexamined over a large number of doses and schedules. The outcomes were examinedexplicitly using a sigmoid-dose-effect model. Once again, T > MIC was the variablemost closely linked to outcome (Table I).

Other groups, such as those of Bakker-Woudenberg (Roosendaal et al., 1986)have also shown that, for drugs such as ceftazidime, T > MIC is the importantdynamically-linked variable in both neutropenic and normal animals infected withKlebsiella pneumoniae (Table II). In this instance, use of continuous infusion resultedin a three- to 15-fold decrease in the total dose of antibiotic necessary to completelyprotect a cohort of animals compared with intermittent administration of the drug.

Most recently, Craig's group (Craig, W., personal communication) provided a mostinteresting insight into this problem. While previous animal model data indicate thatT > MIC is important for /Mactams, the question still arises if there is a certainminimum amount of the dosing interval during which the concentration needs to beabove the MIC. Craig's group addressed this problem by determining the static point,i.e. the percentage of the dosing interval needing to be covered to reach the point wherea bacterial inoculum in an animal thigh neither multiplied nor was killed. Multiple

Dow



Alexander Project: Ptiarmacodynamics 143

-2 12

g

|

20

15

10

0

( b ) j * * * I'* i • I*

11 1 1-2 0 2 3 124 5 6

Time (h)Figure 1. Kinetics of sub-MICs of ticarcillin and the corresponding effect on P. aerugmosa ATCC 27853

in the same granulocytopenic mice, (a) Growth kinetics of P. aeruginosa in ctvo. Each point represents thegeometric mean ± S.D number of cfu per thigh in three mice. The differences among the three growth curvesare significant (P < 0.01). (b) Plasma kinetics after repeated 3 and 1 h subcutaneous injections of ticarcillin(30 and 10/ig/g, respectively). Each point stands for the mean ± S.D plasma level in three mice. Limit ofdetectability ( ). I mg/L. O, Control (saline); A, 30/jg/gq 3 h; • , 10/ig/gq 1 h (cf. Gerber <•/a/., 1983)

dosing intervals were evaluated. Interestingly, for penicillins and cephalosporins, aconstant 50-60% of the dosing interval needed to be covered by drug in order to achievebacterial stasis. This provides an insight into the minimal expectation which we shouldhave of a /?-lactam antibiotic for a specific pathogen. Thus, for empirical use, a /Mactamwith relatively low protein binding should achieve concentrations in excess of the MIG»of the pathogen for which the antibiotic is to be employed for a minimum of 50% ofany proposed dosing interval.

In patients, there are also data that support the concept of T > MIC being linkedto outcome. The data developed by several groups that a serum inhibitory orbactericidal titre of 1:8 or 1:16 at peak antibiotic concentrations (1 h after the end ofan intravenous infusion of 30 min-1 h duration) is linked to outcome supports theconcept that T > MIC is truly linked to outcome (Klastersky et al., 1974; Platt et al.,1981; Sculier& Klastersky, 1984). One doubling dilution in this assay occurs in vivo overa single terminal elimination half life. Consequently, titres of 1:8 or 1:16 indicate that

Dow



144 G. L. Drusano and F. W. Goldstein

Table I. Stepwise multilinear regression analysis of drug efficacy on three predictors(time levels exceeded the MIC, T > MIC; logl0 AUC above the MIC, AUC > MIC; and

logio peak level) (cf. Leggett et al., 1989)

Drug (mg/kg), MIC(mg/L), infection site(no. of mice)

Cefazolin (9.375-1200), 2.4thigh (48)lung (80)

Imipenem (0.586-75), 0.39thigh (36)lung (72)

Ceftazidime (0.293-75), 0.15thigh (52)lung (116)

Gentamicin (1.5-24), 0.27thigh (32)

ql-6h'q6-12h'

lung (80)ql-6h'q6-12h"

Netilmicin (0.1875-24), 0.12thigh (42)

ql-6h'q6-12h'

lung (88)ql-6h'q6-12h'

Parameter(s)selected

T > MIC*T > MIC*

T > MIC*T > MICT > MIC andlog AUC > MIC

T > MIC*T > MIC

T > MICT > MIC andlog AUC > MIClog AUC > MIC*T > MIC*log AUC > MIClog AUC > MIClog AUC > MIC*

T > M I CT > MIC andlog AUC > MIClog AUC > MICT > M I Clog AUC > MIC*log AUC > MIClog AUC > MIC*

R1 (%)

79.668.1

76.876.4

81.9

78.972.9

53.2

61.173.794.145.650.482.8

59.9

69.974.184.281.482.787.5

Coefficient"

-4.21-2.91

-5.11-4.44-3.58-1.02

-5.22-4.06

-3.84-3.04-0.68-1.59-19.3-1.26-1.36-2.01

-5.77-3.80-0.70-1.37-19.8-1.45-2.29-1.50

P

<0.00001< 0.00001

<0.00001<0.00001< 0.00001

0.00001

< 0.00001<0.00001

<0.000010.000010.0111

<0.00001<0.00001<0.00001< 0.00001< 0.00001

<0.00001<0.00001

0.0005<0.00001<0.00001< 0.00001< 0.00001<0.00001

'A larger coefficient indicates a greater change in logio cfu for a given unit of change in theparameter.

*Only parameter selected by stepwise regression (P < 0.05).'Best univariate parameter selected by stepwise regression.'Analysis of regimens with 1-6 h or 6-12 h dosing intervals only; q, every.

three to four terminal half lives can elapse and there will still be sufficient antibioticavailable to inhibit or kill the microorganism. This is 1 h after the end of the infusion(a time span of 1.5-2.0 h). For an antibiotic with a half life of 1 h, this implies aminimum time of 4.5 h of coverage and a maximum time of 6.0 h of coverage. Forantibiotic with a 2 h half life, these figures are 7.5-10 h. As dosing intervals for thesestudies were between 4 and 8 h, it is not surprising these titres were significantlyassociated with a good clinical outcome.

One of the earliest clinical trials which specifically examined this issue was performedby Bodey, Ketchel & Rodriguez (1979). A continuous infusion of cefamandole wascompared with intermittently administered cefamandole. Both groups had intermit-tently administered carbenicillin as part of the regimen. The study examined febrile,

Dow



Tab

le I

I. E

ffic

acy

of

ceft

azid

ime

trea

tmen

t sc

hedu

les*

in

no

rmal

an

d le

ucop

enic

* ra

ts (

cf.

Ro

ose

nd

aal

et a

l.,

1986

)

No

rmal

ra

ts (

n =

10)

inte

rmit

ten

t ad

min

istr

atio

n co

nti

nu

ou

s ad

min

istr

atio

nD

osa

ge

no

. of

ti

me

to

no

. of

ti

me

to(m

g/k

g p

er d

ay)

surv

ivo

rs

dea

th

(day

sy

surv

ivo

rs

dea

th

(day

s)'

Leu

cop

enic

ra

ts (

n =

10

)in

term

itte

nt

adm

inis

trat

ion

con

tin

uo

us

adm

inis

trat

ion

no.

of

tim

e to

n

o.

of

tim

e to

surv

ivo

rs

dea

th

(day

s)'

surv

ivo

rs

dea

th

(day

s^

0.06

0.12

0.23

0.47

0.94

1.88

3.75

7.50

15.0

030

.00

60.0

0

0 1 4 6 9 9 10

8.9

±

2.0

9.5

±3

.18.

2 ±

3.

38.

8 ±

1.

5

0 2 1 6 10

5.5

±0

.75.

8 ±

0.7

9.0

± 1

.46.

8 ±

1.6

0 1 7 10

3.4

±

1.2

3.5

±0

.75.

3 ±

2.1

0 1 7 10

2.8

±

0.4

5.2

±

1.9

9.3

±3

.5

"To

grou

ps o

f te

n ra

ts e

ach

ceft

azid

ime

was

adm

inis

tere

d ei

ther

as

inte

rmit

tent

bol

us i

njec

tions

at

6 h

inte

rval

s o

r as

a c

onti

nuou

s in

fusi

on (

infu

sion

rat

e, 0

.113

ml/

h)ov

er a

per

iod

of 4

day

s. T

reat

men

t w

as s

tart

ed

5h

afte

r in

ocul

atio

n of

th

e le

ft

lung

with

8 x

10*

cfu

of

K.

pneu

mon

iae.

T

he

PD

»s (

mg/

kg/d

ay)

for

inte

rmit

tent

adm

inis

trat

ion

and

cont

inuo

us a

dmin

istr

atio

n fo

r no

rmal

rat

s w

ere

0.35

an

d 0.

36,

resp

ectiv

ely

(99.

9% c

onfi

denc

e li

mit

s, 0

.19-

0.67

an

d 0.

21-0

.61,

res

pect

ivel

y);

the

corr

espo

ndin

g va

lues

for

leu

cope

nic

rats

wer

e 24

.37

and

1.52

(16

.07-

36.9

7 an

d 1.

00-2

.31)

.te

ftaz

idim

e w

as a

dmin

iste

red

intr

aper

iton

eall

y in

tw

o do

ses

of 9

0 an

d 60

mg/

kg a

t 5

days

an

d 1

day

bef

ore

bact

eria

l in

ocul

atio

n,

resp

ectiv

ely.

'Mea

n ±

S.D

., ba

sed

on

the

tim

e of

bac

teri

al i

nocu

lati

on (

day

0).

x I i "3

Dow




Table III. Continuous versus intermittent administration of cefamandole. Each pluscarbenicillin in febrile cancer patients (cf. Bodey el al., 1979)

Patientgroup

(1) Initial neutrophilcount < 100

(2) Profound, persistentneutropenia

(3) Causative organismcefamandole sensitive

Preferredtreatment group

continuousinfusion

continuousinfusion

continuousinfusion

Response andstatistical difference

71% vsO%P = 0.06

65% D5 21%P = 0.03

92% vs 63%/> = 0.04

neutropenic cancer patients. Table III demonstrates that the outcomes always tendedto be better with continuous infusion and were significantly better in two of the patientgroups. Warren et al. (1983) examined cefoperazone as part of a randomized evaluationin septic patients. Ten patients had documented single organism, Gram-negativebacillus bacteraemia. Nine of these patients had a correct prediction of success or failuremade on the basis of the expected time above the MIC.

Perhaps the most convincing clinical evidence comes from the study by Schentag et al.(1984). These investigators examined the cephalosporin, cefmenoxime, in patients withGram-negative pneumonia. They documented the time that plasma concentrationsexceeded the MIC of the pathogen (T > MIC) and also measured the time toeradication of the pathogen from the lower tracheobronchial tree. There was anexcellent inverse relationship between T > MIC and time to bacterial eradication(Figure 2).

Consequently, there are multiple lines of evidence (in-vitro, animal model andclinical), suggesting that T > MIC is the pharmacological variable most closely linkedto bacterial killing and a good clinical outcome for the /J-lactam class of agents. It isalso likely that only the free drug is microbiologically active (Drusano, 1988) and,consequently, a conservative evaluation of this factor (T > MIC) would employ the freedrug in this calculation.

The other factor which affects our judgement regarding the potential utility of a/3-lactam antibiotic is the MIC for the organism(s) against which the drug is to be usedempirically. The reason for examining the distribution of the MICs is quitestraightforward. While T > MIC is the variable which is truly linked to outcome, mostof the change in this variable is a consequence of changes in the MIC. For example,a two-fold range of dose may be available for a licensed agent (e.g. 250 or 500 mg).There may be a 50% coefficient of variation in the clearance of the drug in thepopulation of patients in which use is intended; this will produce (approximately) asix-fold range of exposure (the pure pharmacological variable). However, there is oftena 64- or even a 128-fold range in the MICs for the pathogens for which the drug isintended for empirical use. Consequently, there is a much greater degree of variabilityin the microbiological end of the equation relative to the pharmacological end.Therefore, the determination of the distribution of MICs of a large number of drugsfor respiratory pathogens in different areas of the world is a critically important pieceof information if these agents (and in particular, the /Mactam drugs) are to be placedin proper perspective.

Dow



Alexander Project: Pharmacodynainics 147

It is one of the aims of the Alexander Project to generate a database with a substantialnumber of isolates derived from Western Europe and the USA and to determine theirsusceptibilities to a panel of antimicrobial agents employing a standardizedmethodology. This database is one which is being developed, allowing trends insusceptibility for multiple antibiotics to be recognised, by site, by time and always withthe aim of placing the potential utility of these commonly used agents in properperspective. This can be done most straightforwardly and rationally for the /?-lactamsby examining the time that antibiotic concentrations exceed the MIC90 of these clinicallyimportant pathogens; the most conservative estimate is to use the free drugconcentration in any analysis.

How does one assess the other classes of antibiotics under investigation in theAlexander Project, and in particular macrolides and fluoroquinolones, which are widelyused in the participating countries?

Fluoroquinolones exhibit concentration-dependent bactericidal activity, and thera-peutic outcome should be linked to AUC. This was clearly demonstrated by Leggettet al. (1991) for ciprofloxacin in murine pneumonitis and thigh infection models.

The situation for macrolides is confused, partly because erythromycin is generallyconsidered as a bacteriostatic agent, but bactericidal against streptococci.Clarithromycin and azithromyan are considered more bactericidal than erythromycin(Piscitelli, Danziger & Rodvold, 1992). In addition, macrolides are concentrated bymacrophages and azithromycin is characterised by slow efflux from cells (ion-trapping).Gerber (1990) noted that, for a range of pathogens in the neutropenic mouse thigh

0 2 4 6 8 10Cefmenaxime above DRC in vivo (h)

Figure 2. Relationship between the time cefmenoxime serum concentrations exceeded the dynamic responseconcentration (DRC) for retrospectively (A) and prospectively (#) treated patients, and the days to bacterialeradication in vivo. Each data point represents one pathogen. The regression line describing the retrospectivedata took the form of the equation: days to eradication = 13.86-1.78 (time over dynamic responseconcentration), r = 0.89, P < 0.001. The dual individualized dosage method clustered the eradication daytightly around days 4 to 6. (cf. Schentag el al., 1984).

Dow




Table IV. Pharmacodynamic summary of Alexander Project data for antibiotics against5. pneumoniae showing concentration-independent (a) and -dependent (b) killing

(a)Antibiotic

Amoxycillin

Amoxycillin/clavulanate

CefaclorCefuroxime

Cefixime

CeftriaxoneErythromycinClarithromycin

(b)Antibiotic

ClarithromycinAzithromycinCiprofloxacin

Ofloxacin

Dose (mg) Regimen

250500250500250250500200400

1000500250

Dose (mg)

250500500750400

tid

tid

tidbd

bdododqidbd

M I C M I C(mg/L) (mg/L)

0.06 20.06 81 2

2 2

M I C(mg/L) T :

0.015

0.015

0.50.03

0.25 :

0.030.060.06 :

AUC

> MIC

6.8>8

6.8>8

3.411.9

> 12> 12

22.3>24> 6

>12

(mg.L.h)o.24h

6.1 (4.3Y2.4

15.524.247.5

MICV(h) (mg/L)

1

1

644

16

142

RatioAUC/MIC

101.7(173.3)'4015.524.223.8

T > M I C ( h )

3.54.73.54.700300

>2400

RatioAUC/MIC

3.1 (5.2Y0.37.8

12.123.8

•T > MIC, Time above MIC (h) for one dosing period, it multiply time by number of doses per day foroverall coverage/24 h.

*AUCs estimated by trapezoidal rule.'14-OH darithromycin.

infection model, the parameter most closely associated with efficacy for erythromycinwas T > MIC. For azithromycin, however, AUC correlates best with outcome, thereason being that the drug concentrates into white cells which then act as a deliverysystem to the site of infection (Craig, W., personal communication). For clarithromycin,both T > MIC and AUC are linked to outcome, but it is not yet possible to determinewhich is the more important. Thus, we can consider erythromycin in the same way as/?-lactams, azithromycin in the same way as quinolones, and clarithromycin in eitherscheme.

However, it should be noted that the ratio of serum concentrations to MIC forquinolones can be important in terms of development of resistance. For amoxycillin andfluoroquinolones, resistant mutants can be selected if the serum level is <4 x MIC andfor cephalosporins, selection occurs at < 16 x MIC. As shown in the present study, theratio of serum level:MIC for amoxycillin against Streptococcus pneumoniae is in therange 8-1000, while for fluoroquinolones, cefaclor and cefixime the ratios range from2 to 16. Thus, there is potential for resistance development on treatment with theseagents.

Since there is no unifying scheme to compare all classes of antibiotics, how can thesepharmacodynamic parameters be compared using the MIC data generated by theAlexander Project? Tables IV to VII summarise the pharmacokinetic parameters of theantibiotics tested using the recommended dosage(s), the MIC50 and MIG» against the

Dow



Alexander Project: Ptiarmacodynamics 149

four main pathogens and T > MIC or AUC/MIC values (as appropriate), calculatedfrom serum/time profiles. Figures 3 to 6 show the proportion of the dose interval that/3-lactam and macrolide serum levels exceed the MIC. The following conclusions canbe drawn:

Antibiotic class

(i) Of the penicillins, only amoxycillin/clavulanate (at the higher dose regimen)exceeds the MIG» for the four major pathogens for > 50% of the dosing interval.Amoxycillin (higher dose) was as effective as amoxycillin/clavulanate against S.pneumoniae, but not against the other pathogens.

(ii) Of the cephalosporins, only ceftriaxone exceeds the MIQo for the four majorpathogens for > 50% of the dosing interval. Cefixime was effective against Haemophilusinfluenzae and Moraxella catarrhalis, but none of the other oral cephalosporinsexceeded the MIG» for any organism for > 50% of the dose interval.

(iii) Macrolides showed efficacy against M. cattarhalis only (at MIC™ values).(iv) Quinolones showed the greatest potential efficacy against H. influenzae and

M. catarrhalis, and the least activity against Slaphylococcus aureus and S. pneumoniae(at MIC* values).

Table V. Pharmacodynamic summary of Alexander Project data for H. influenzae for drugsshowing concentration-independent (a) and -dependent (b) killing

(a)Antibiotic Dose (mg) Regimen

MIC*, MIC*(mg/L) T > MlCjo'tfi) (mg/L) T > MICV(h)

Amoxycillin


CefaclorCefuroxime

Cefixime


250500250500250250500200400

1000500250

tid

tid

tidbd

bdododqidbd

0.5

0.5

41

0.12

0.1244

4.85.64.85.60.63.75.9

>12>24>24

00

16

1

82

0.12

0.1248

003.64.602.54.5

>12>24>24

00

(b)Antibiotic Dose (mg)

M I C(mg/L)

MIC*(mg/L)

AUC Ratio RatioA U C / M I C M AUC/MIC*


Ofloxacin

250500500750400

40.50.03

0.03

810.03

0.03

6.1 (4.3)'2.4

15.524.247.5

1.5(2.6)'4.8517807

1583

0.76(1.3)'2.4

517807

1583

•T > MIC = Time above MIC (h) for one dosing period, ie multiply time by number of doses per day foroverall coverage/24 h.

'AUCs estimated by trapezoidal rule.'14-OH clarithromycin.

Dow




Table VI. Pharmacodynamic summary of Alexander Project data for M. catarrhalis for drugsshowing concentration-independent (a) and -dependent (b) killing

(a)Antibiotic

Amoxycillin

Amoxycillin/Clavulanate

CefaclorCefuroxime

Cefixime


(b)Antibiotic


Ofloxacin

Dose (mg) Regimen

250500250500250250500200400

1000500250

Dose (mg)

250500500750400

MIC»(mg/L)

0.060.060.03

0.12

tid

tid

tidbd

bdododqidbd

MIG»

0.120.060.06

0.12

M I C

0.5

0.06

0.51

0.12

0.120.120.06

M I C, T>MIC»'(h) (mg/L)

4.85.6

>8>8

3.43.75.9

>12>24>24

3.7>12

A.UC'(mg.L.h)o-24h

6. 1 (4.3X2.4

15.524.247.5

4

0.25

12

0.5

10.120.12

RatioAUC/MIOo

102(173y40

517807396

T > MICV(h)

12.45.87.12.62.54.56.4

18.2>24

3.711

RatioAUC/MIC«

51 (Siy40

258403396

•T > MIC, Time above MIC (h) for one dosing period, ie multiply time by number of doses per day foroverall coverage/24 h.

*AUCs estimated by trapezoidal ruler14-OH clarithromycin.

Pathogen

(i) For S. pneumoniae, all /?-lactams except cefaclor had serum levels above thewhich is close to the MIC for normally susceptible strains, for > 50% of the doseinterval. This is also true for macrolides, and to a lesser extent for ofloxacin andhigher-dose ciprofloxacin.

At the MIC90, which corresponds to the modal MIC for resistant strains, onlyceftriaxone and amoxycillin (± clavulanate) of the 0-lactams showed T > MIC90exceeding 50% of the dose interval. Cefaclor, cefuroxime and cefixime are out oftherapeutic range for these particular isolates. Ofloxacin had the highest AUC/MIC90ratio.

(ii) For H. influenzae, at the MIC», ceftriaxone, cefixime and the fluoroquinolones,followed by amoxycillin (± clavulanate) have excellent activity. Cefuroxime hasborderline activity, with possible usefulness against some strains. Macrolides areineffective against these strains since serum levels are lower than MICs. At the MIG»,amoxycillin is also less effective because of /Mactamase production.

(iii) For M. catarrhalis, at the MIC», amoxycillin (± clavulanate), cefixime,ceftriaxone, macrolides and fluoroquinolones are potentially effective. Cefaclor andcefuroxime may be useful against some strains. At the MICw, only cefixime, ceftriaxone

Dow



Alexander Project: Pharmacodynamlcs 151

Table VII. Pharmacodynamic summary of Alexander Project data for 5. aureus for drugs showingconcentration-independent (a) and -dependent (b) killing

(a)Antibiotic

Amoxycillin


CefaclorCefuroxime

Cefixime


(b)Antibiotic


Ofloxacin

Dose (mg) Regimen

250500250500250250500200400

1000500250

Dose (mg)

250500500750400

M I C(mg/L)

0.120.50.25

0.25

tid

tid

tidbd

bdododqidbd

MIC»(mg/L)

3232

1

2

MIC*,(mg/L) T :

1

0.5

21

16

2 :0.250.12

AUC(mg.L.h),

>MIC

3.54.74.85.61.53.75.900

>24>611

6.1 (4.3)f

2.415.524.247.5

MIC*,V(h) (mg/L)

32

1

82

32

43232

RatioAUC/MIC»

51 (87y4.8

6297

190

T > MICV(h)

003.54.702.54.500

>2400

RatioAUC/MIC

0.2 (0.33^0.08

15.524.224

•T > MIC, Time above MIC (h) for one dosing period, ie multiply time by number of doses per day foroverall coverage/24 h.

*AUCs estimated by trapezoidal rule.M4-OH clarithromycin.

Amoxycillin 250Amoxycillin 500

Amoxycillin/clavulanate 250/125Amoxycillin/clavulanate 500/125

Cefaclor 250

Cefuroxime 250Cefuroxime 500

Cefixime 200Cefixime 400

Ceftriaxone 1 g

Erythromycin 500

Clarithromycin 250

0 10025 50 75% of dose interval above MJC

Fignre 3. Time > MIC as percentage of dosing interval for S. pneumoniae. • , MIC», • . MIOo

Dow






Cefaclor 250


Cefixime200Cefixime400

Ceftriaxone 1 g

Erythromycin 500

Clarithromycin 250

0 25 50 75 100% of dose interval above MIC

Figure 4. Time > MIC as percentage of dosing interval for H. influenzae. • , MIC»; • , MIG».

and amoxycillin/clavulanate of the /Mactams, the macrolides and quinolones are likelyto be effective.

(iv) For S. aureus, at the MIC50, ceftriaxone, amoxycillin/clavulanate, macrolides,fluoroquinolones, amoxycillin (higher dose) and cefuroxime (higher dose) are likely tobe effective. Cefaclor may be useful against some strains, but cefixime showed pooractivity. At the MIC90, only ceftriaxone and amoxycillin/clavulanate (higher dose) arelikely to be effective. High dosage ciprofloxacin, ofloxacin and cefuroxime may be usefulagainst some strains, but other /Mactams and macrolides have little to offer.



Cefaclor 250


Cefixime 200Cefixime 400

Ceftriaxone 1 g

Erythromycin 500

Clarithromycin 250

25 50 75% of dose interval above MIC

100

Figure 5. Time > MIC as percentage of dosing interval for hi. caiarrhalis. • , MIC»; • , MIC«

Dow



Alexander Project: Pharmacodynamlcs 153

Amoxydllin 250Amoxycillin 500


Cefador 250

Cefuroxime 250Cefuroiime 500

Ce&rime 200Cefixiine400

Ceftriaxone 1 g

Erythromycin 500

Clarithromycin 250

0 25 50 75% of dose interval above MIC

100

Figure 6. Time > MIC as percentage of dosing interval for S. aureus. • , MIC»; • , MICV

Conclusion

How do we interpret these data? All the parameters have been calculated on the basisof the usual mean unitary dosage. For amoxycillin, because of lack of toxicity, dosagescan be increased by 100%. As an example, the recommended oral dosage of amoxycillinfor pneumococcal pneumonia in France is 1 g tid, compared with 500 mg tid in manyother countries. This is not normally done for other oral /Mactams. Protein bindingshould also be considered, since this may seriously compromise the activity of somecompounds, such as cefixime (protein binding = 70%) against S. pneumoniae. Selectionof antibiotic should be made on the basis of the site of infection, the need for bactericidalactivity and the toxicity profile, allied to a knowledge of the pharmacodynamics, forwhich good susceptibility data as provided by the Alexander Project are of paramountimportance.

References

Bodey, G. P., Ketchel, S. J. & Rodriguez, V. (1979). A randomised study of carbenicillin pluscefamandole or tobramycin in the treatment of febrile episodes in cancer patients. AmericanJournal of Medicine 67, 608-16.

Drusano, G. L. (1988). Role of pharmacokinetics in the outcome of infections. AntimicrobialAgents and Chemotherapy 32, 289-97.

Gerber, A. U. (1990). Impact of the antibiotic dosage schedule on efficacy in experimental softtissue infections. Scandinavian Journal of Infectious Diseases 74, Suppl., 147—54.

Gerber, A. U., Craig, W. A., Brugger, H.-P., Feller, C , Vastola, A. P. & Brandel, J. (1983).Impact of dosing intervals on activity of gentamicin and ticarcillin against Pseudomonasaeruginosa in granulocytopenic mice. Journal of Infectious Diseases 147, 910-7.

Klastersky, J., Daneau, D., Swings, G. & Weerts, D. (1974). Antibacterial activity in serum andurine as a therapeutic guide in bacterial infections. Journal of Infectious Diseases 129, 187-93.

Leggett, J. E., Ebert, S., Fantin, B. & Craig, W. A. (1991). Comparative dose-effect relations atseveral dosing intervals for beta-lactams, aminoglycoside and quinolone antibiotics againstgram-negative bacilli in murine thigh-infection and pneumonitis models. ScandinavianJournal of Infectious Diseases 74, Suppl., 179-84.

Dow



154 G. L. Dnisano and F. W. Goldstein

Leggett, J. E., Fantin, B., Ebert, S., Totsuka, K., Vogelman, B., Calame, W. el al. (1989).Comparative antibiotic dose-effect relations at several dosing intervals in murinepneumonitis and thigh-infection models. Journal of Infectious Diseases 159, 281-92.

Piscitelli, S. C , Danziger, L. H. & Rodvold, K. A. (1992). Clarithromycin and azithromycin: newmacrolide antibiotics. Clinical Pharmacy 11, 137-52.

Platt, R., Erlich, S. L., Afarian, J., O'Brien, T. F., Pennington, J. E. & Kass, E. H. (1981).Moxalactam therapy of infections caused by cephalothin-resistant bacteria: influence ofserum inhibitory activity on clinical response and acquisition of antibiotic resistance duringtherapy. Antimicrobial Agents and Chemotherapy 20, 351-5.

Roosendaal, R., Bakker-Woudenberg, I. A. J. M., van den Berge-van Raffe, M. & Michel, M. F.(1986). Continuous versus intermittent administration of ceftazidime in experimentalKlebsiella pneumoniae pneumonia in normal and leukopenic rats. Antimicrobial Agents andChemotherapy 30, 403-8.

Schentag, J. J., Smith, I. L., Swanson, D. J., E>e Angelis, C , Fracasso, J. E., Vari, A. el al. (1984).Role for dual individualization with cefmenoxime. American Journal of Medicine 71, Suppl.,6A, 43-50.

Sculier, J. P. and Klastersky, J. (1984). Significance of serum bactericidal activity inGram-negative bacillary bacteremia in patients with and without granulocytopenia.American Journal of Medicine 76, 429—35.

Warren, J. W., Miller, E. H., Fitzpatrick, B., DiFranco, D. E., Caplan, E. S., Tenney, J. H. et al.(1983). A randomised, controlled trial of cefoperazone vs cefamandole-tobramycin in thetreatment of putative, severe infections with Gram-negative bacilli. Reviews of InfectiousDiseases 5, Suppl. /, SI73-80.

Zinner, S. H., Dudley, M. N., Gilbert, D. & Bassignani, M. (1988). Effect of dose and scheduleon cefoperazone pharmacodynamics in an in vitro model of infection in a neutropenic host.American Journal of Medicine 85, Suppl. IA, 56-8.

Dow



relevance of the alexander project

Documents