Transcript 08-Drusano

The Endpoint from a Resistance Point of View
A Symposium to Honor the Career of
William A. Craig, M.D.
George Drusano, M.D.
Co-Director
Ordway Research Institute
A Colleague Responsible for Much
of What I am About to Show
Dr. Arnold Louie
A Relevant Quote
All Models Are Wrong;
Some Models Are Useful
Professor G.E.P. Box, a famous and very
wise biostatistician
Pharmacodynamics
• Let us look at an example of “classical
antibacterial pharmacodynamics”, where a
survivorship endpoint was employed
Role of PK/PD – Survivorship
Endpoint
• In this P. aeruginosa rat sepsis
model, where a
fluoroquinolone was employed
for therapy (an AUC/MIC Ratio
drug), several things are
evident:
1. As MIC increases,
survivorship decreases (same
dose – isogenic mutants)
2. As Dose increases (same
MIC) survivorship increases
3. Equal AUC/MIC ratios
produce equivalent outcomes
Drusano et al. Antimicrob Agents Chemother
1993;37:483-490
Pharmacodynamics
• One also needs to think about other
desired endpoints:
1. clinical outcome/survivorship
2. microbiological outcome/cell kill
3. suppression of resistance
4. organism population eradication (TB)
• The amount of drug exposure to
achieve the endpoints does differ
• Let us look at resistance suppression
Resistance Suppression and
Outcome Optimization
• First, we will examine Pseudomonas aeruginosa
in a hollow fiber infection model
• Then we will examine this same pathogen in a
mouse thigh model of infection
• Then we will examine Staphylococcus aureus
and:
1. examine the impact of therapy duration
2. examine the divergent effect on sensitive and
resistant subpopulations when drug is withdrawn
3. examine the impact of granulocytes
Resistance Suppression and Outcome
Optimization
Journal of Clinical Investigation 2003;112:275-285 &
Nature Reviews Microbiology 2004;2:289-300
Resistance Suppression and Outcome
Optimization
The use of the hollow fiber model for studying antimicrobial regimens was described
by Blaser and Zinner and employed extensively by Dudley
Infusion
+
Central
Compartment (Cc)
Bacteria
(XT/R)
SCl
f(c)
H
dCc=Infusion-(SCl/V)xCc dXS=KGS x XS x L - fKS(Cc ) x XS
[1]
dt
dt
dXR= KGR x XR x L- fKR(CcH ) x XR
[2]
[3]
dt
L = (1-(XR
+ XS)/POPMAX)
Kmax  CcH 
[4]
, =K and  = S,R
[5]
Y1=XT=XS+XR, IC(1)= Initial Total Population
[6]
Y2=XR , IC(2)= Initial Resistant Population
[7]
f(CcH)=
C H  50 +CcH 
Resistance Suppression and Outcome
Optimization
Tam V et al. Bacterial-population responses to drug selective pressure: Examination
of garenoxacin’s effect on Pseudomonas aeruginosa. J Infect Dis 2005;192:420-428
Resistance Suppression and Outcome Optimization
• The amplification of the
resistant sub-population
is a function of the
AUC/MIC ratio
• The response curve is an
inverted “U”.
• The AUC/MIC ratio for
resistant organism stasis
is circa 185/1
Resistant organisms
at baseline
Resistant Mutants (CFU/mL)
P. aeruginosa - Prevention of Amplification of Resistant
Subpopulation
107
106
105
104
103
100
10
0
50
100 150 200 250
AUC0-24:MIC Ratio
All other data points represent
resistant organism counts at
48 hours of therapy
Resistance Suppression and Outcome
Optimization
Propspective Validation Study
Tam V et al. Bacterial-population responses to drug selective pressure: Examination of
garenoxacin’s effect on Pseudomonas aeruginosa. J Infect Dis 2005;192:420-428
Levofloxacin and Pseudomonas aeruginosa
in a Mouse Thigh Infection Model
Can a drug exposure be
identified that will prevent the
resistant subpopulation from
taking over the total
population?
In contrast to the hollow fiber
system, here we have an
immune system
P. aeruginosa outcome
studies
Jumbe et al J Clin Invest 2003;112:275-285
90 mg/kg
0 mg/kg
215 mg/kg
Journal of Clinical Investigation 2003;112:275-285 &
Nature Reviews Microbiology 2004;2:289-300
600 mg/kg
Peripheral (thigh)
Compartment (Cp)
Central Blood
Compartment (Cc)
ke
[1]
[2]
[3]
dCa= -kaCa
dt
dCc= kaCa+kpcCp-kcpCc-keCc
dt
dCp = kcpCc - kpc Cp
dt
+
kpc
kcp
IP
injection
Bacteria
(XT/R)
f(c)
dXS=KGS x XS x L - fKS(CcH ) x XS
dt
dXR= KGR x XR x L- fKR(CcH ) x XR
dt
[4]
[5]
L = (1- (XR + XS)/POPMAX)
f(CcH)=
Kmax  CcH 
, =K and  = S,R
[6]
[7]
C H 50+CcH 
Y1=XT=XS+XR
[8]
Y2=XR
[9]
Jumbe et al J Clin Invest 2003;112:275-285
Drusano GL. Nat Rev Microbiol 2004;2:289-300
Resistance Suppression and Outcome
Optimization
Jumbe et al J Clin Invest 2003;112:275-285
Drusano GL. Nat Rev Microbiol 2004;2:289-300
Journal of Clinical Investigation 2003;112:275-285 &
Nature Reviews Microbiology 2004;2:289-300
• We studied the quinolone
garenoxacin against S.
aureus
• We studied and modeled
7 regimens
• We performed a
prospective validation
with 4 hypotheses
• All were validated
• Resistance suppression
requires more drug
exposure than that for
maximal rate of kill
PK/PD
PK/PD
• We performed an experiment where 4, 5 or 6
daily doses of drug (AUC/MIC ratio=100) were
administered and the outcomes monitored out
to day 13
• We fit an expanded mathematical model to all
the data simultaneously
• We included a natural death rate term for
sensitive and resistant populations
• This allows us to look at relative biofitness in
conjunction with the growth terms for the two
populations
PK/PD
PK/PD
PK/PD
• So, the longer therapy continues, the more
amplification goes on of the resistant
population with a suboptimal regimen
• To prevent resistance, shorter is better
• BUT, we also have to clear the infection, so
therapy needs to be long enough to
accomplish this end
• How much effect do we need?
• Enough to allow the immune system to do its
job!
Parameter
Units
Kmax-growth
h-1
POPMAX
CFU/g
Vmax-kill *
h-1
Km
CFU/g
Mean
0.622
0.916x1011
0.00535
147660
SD
0.104
0.980x1010
0.00400
203928
______________________________________________________________________
* Rate constant is multiplied by the granulocyte count
BRIDGING TO THE CLINIC
Correlation with Pre-Clinical Models
Do These Animal and Hollow Fiber Model
Findings Correlate with Suppression of
Resistance in the Clinic?
PK-PD TARTGET ATTAINMENT
Ciprofloxacin Against P. aeruginosa
Use of Monte Carlo Simulation
Taking the
expectation
demonstrates an
overall target
attainment of
62% and a
predicted
emergence of
resistance rate of
38%
MONTE CARLO SIMUATION
Is It Predictive?
• Peloquin studied 200 mg IV Q 12 h of ciprofloxacin in
nosocomial pneumonia - P aeruginosa resistance rate 70% (7/10
- pneumonia only) - 77% (10/13 - all respiratory tract)
• MCS (resistance suppression target) predicts emergence of
resistance in 75%
• Fink et al studied ciprofloxacin in nosocomial pneumonia (400
mg IV Q 8 h) - P aeruginosa resistance rate 33% (12/36)
• MCS at this dose and schedule predicts suppression in 62% and
emergence of resistance in 38%
Peloquin et al Arch Int Med 1989;1492269-73
Fink et al AAC 1994;38:547-57
Sometimes, single agent therapy just
can’t get the job done
WHAT ABOUT COMBINATION
THERAPY FOR RESISTANCE
SUPPRESSION?
Cefto-Levo Combo vs. P
aeruginosa
• We wished to evaluate the combination of
ceftobiprole plus levofloxacin
• West et al (Clin Ther 2003;25:485-506)
demonstrated that levofloxacin in combination
with a β-lactam suppressed Pseudomonas
resistance in 17/17 instances in a HAP trial
• We employed a mouse thigh infection model to
examine the cefto-levo combination for both cell
kill and resistance suppression
Cefto-Levo Combo vs. P aeruginosa
Methods
• The mouse thigh infection model as
described by Craig was employed (inoculum
1.85 x 107 CFU)
• P aeruginosa ATCC 27853 was employed as
the challenge organism (MIC 1.0 – Levo;
MIC 2.0 mg/L – Cefto)
• Dose ranging studies were performed for
each drug separately
• These were used to design the combination
experiment using D-optimal design theory
Cefto-Levo Combo vs. P aeruginosa
Methods
• Drug concentrations were measured by LC-MS/MS
techniques
• Resistant organisms were determined by plating on
agar infused with 2.5XMIC (Levo) or 3XMIC (Cefto)
• Combination cell kill was modeled employing the
Universal Response Surface Approach of Greco et al
• The ability to suppress resistant subpopulations was
evaluated by Monte Carlo Simulation (Lodise et al
AAC 2007;51:2378-2387 for Ceftobiprole; Drusano
et al JID 2004;189:1590-1597 for levofloxacin)
Cefto-Levo Combo vs. P aeruginosa
Results
Ceftobiprole Single Drug
Therapy
Levofloxacin Single drug
Therapy
Ceftobiprole-Levofloxacin
Combination Therapy
Suppress Levofloxacin
Resistance
Suppress Ceftobiprole
Resistance
Ceftobiprole-Levofloxacin
Combination Therapy
Suppress Ceftobiprole
Resistance
Suppress Levofloxacin
Resistance
Ceftobiprole-Levofloxacin
Combination Therapy
• The exposures that suppress resistance
in both directions are Ceftobiprole T >
MIC of 40% and Levofloxacin AUC/MIC
of 6.7
• Monte Carlo simulation demonstrates
that doses of 750 mg daily of
levofloxacin plus 500 mg of ceftobiprole
as a 2 hour infusion every 8 hours
achieve these targets 96.4% of the time
at a minimum
Ceftobiprole-Levofloxacin
Combination Therapy
• Ceftobiprole plus levofloxacin interact
productively against Pseudomonas aeruginosa
in the mouse thigh model
• Perhaps more importantly, resistance
suppression is clearly demonstrable at modest
drug exposures for both ceftobiprole and
levofloxacin
• Human PK demonstrates that these resistance
suppression exposures can be achieved
greater that 96% of the time
Ceftobiprole-Levofloxacin
Combination Therapy
• THIS MAKES NO SENSE
• For both levo and cefto, the Mex pumps cause
efflux
• This means that there should be no cross
protection – Right??????
• BUT pumps are Michelis-Menten processes –
perhaps the combination saturates the pump
• So, this is a promising combination, at least for
wild-type strain
Anti-Infective PK/PD
I have been trying to interest the anti-infective
community in antimicrobial pharmacodynamics
for almost a quarter of a century, certainly
without notable success.
WELL!
GeorgeGeorge
THANK YOU
FOR YOUR ATTENTION