Transcript Okusanya-ISAP-Post-I..

A Pharmacodynamic Model for Cefprozil
against Haemphilus influenzae in an in vitro
Infection Model across Multiple Regimens
Olanrewaju O. Okusanya, Pharm.D, BCPS
University at Buffalo, School of Pharmacy and
Pharmaceutical Sciences
Co-Authors
Alan Forrest, Pharm.D
Brent M. Booker, Pharm.D
Pamela Kelchin,BS
Patrick F. Smith, Pharm.D
University at Buffalo, School of Pharmacy and
Pharmaceutical Sciences, Buffalo, NY
Introduction/Objectives
•
Adequately characterizing the pharmacodynamics of β-lactams can
give insight into avenues to optimizing the use of this class of drug and
reduce the incidence of resistance development
•
Serial samples of bacterial CFU to track the time course of drug effect
is difficult to obtain in human trials.
•
Modeling concentration vs. rate and extent of kill and/or re-growth
using in vitro PK/PD models (IVM) can provide important insights into
concentrations needed to optimize outcomes in man
•
We have used a mathematical pharmacodynamic model to
characterize bacterial rates of replication and death and the effect of
cefprozil on these processes when administered using different
regimens
Methods
•
•
•
•
•
•
MICs of 2 β-lactamase (-), and 1 β-lactamase (+) H.flu strains were
determined in triplicate following NCCLS criteria
A 1 compartment IVM consisting of a central compartment containing
bacteria in Log-phase growth ( 107 CFU/mL) and bacterial growth
media was used for the experiment
Drug free media was delivered into the central compartment using a
high precision computer driven pump and bacterial growth media
removed to a waste flask
Bacteria were exposed to cefprozil administered into the central
compartment with changing concentrations consistent with that seen in
humans, similar free drug concentrations, with an approximate half-life
of 2 hrs
The once daily (QD), twice daily (BID), and continuous infusion (CI)
regimens were simulated to obtain %T>MIC ranging from 6.6% 100+% with appropriate growth control for each strain and regimen
Serial samples were obtained over 24 hours to determine bacterial
CFU in duplicate with a lower limit of detection of 102 CFU/mL
Pharmacodynamic Model
H

 Conc  
 Imax  
 
MIC  


Effect  1 
H

 Conc  
H
SIT


 
Mi

 MIC  

•CFUi - CFU/mL of the ith population
•VGmax - maximum velocity
of growth (CFU/mL/hr)
(-)
•CFUmi - CFU/mL associated with
½ maximal growth
•CFUTOT - sum of all subpopulations Replication
•Kd – 1st order rate constant
for bacterial death (hr-1)
EC5
0
Drug
Bacteria
CFU/mL
Pop 1
Pop2
Pop3
Pop4
(+)
KD
•Imax - maximum effect
•(Conc/MIC) - approx inverse
serum inhibitory titer (SIT)
•H is Hill’s constant,
•SITmi is the SIT associated
with 50% of maximum drug
effect
EC50
dCFUI
VGmax  CFUI

 Kdo  CFUI
dt
CFUmi  CFUTOT
[ Net Growth ] = [Bacteria replication ] - [ Bacteria death ]
Figure 1: Pharmacodynamic model. CFU/mL, bacteria colony forming units/mL;
Pop1-4, bacteria sub-populations; KD, first-order bacterial death rate constant (hr-1);EC50,
drug concentration of 50% inhibition of growth or acceleration of death (µg/mL)
Methods
• Cefprozil was modeled as either enhancing Kd or inhibiting
replication
• Each curve was fit individually and later simultaneously
regardless of regimen for each strain using maximum likelihood
and then MAP Bayesian estimation
• Each inoculum modeled was allowed to have up to 4 subpopulations
• AIC, Sum of Squares residuals, and visual inspection was used
for model discrimination, to determine the number of
subpopulations, and parameters that would be allowed to vary
between sub-populations
Results
Strain X219
CI regimen
Strain X219
BID regimen
Strain X208
QD regimen
Strain X208
BID regimen
r2=0.94
r2=0.94
Obs=1.00x Fit-0.0
Obs=1.01x Fit-0.23
Strain X219
Strain X208
• CI regimen
• QD regimen
• BID regimen
• BID regimen
r2=0.89
Obs=1.00x Fit-0.28
Strain X204
•QD regimen
Results
• The MICs of the β-lactamase (+) strain (X208) was 2 mg/L and
was 2 mg/L and 4 mg/L for the β-lactamase (-) strain (X219 &
X204)
• Drug effect was best modeled as inhibiting replication (up to
100%)
• The model required 1-2 sub-populations with the VGmax
allowed to vary between sub-populations
• The fit of the pharmacodynamic models for each strain was
excellent accommodating all regimens studied
Results
Parameter
X204
Average VGmax (CFU/mL/hr) 9.02x108
CFUm (CFU/mL)
3.89x108
X219
X208
1.18x109
4.95x108
2.43x108
1.11
T ½ d (hr)
1.00
8.00x108
1.53
cVGmax
--
0.709
0.694
Imax
Hill’s Constant
1.00
2.00
0.992
2.17
0.999
1.00
SITm1
% POP 1
0.361
100
0.612
99.6
1.78
99.9
SITm2
% POP 2
--
115
636
--
0.371
0.061
VGmax: maximum velocity of growth (CFU/ml/hr), CFUm: CFU/mL associated with half the maximal growth, T1/2d: half
life of bacteria giving no replication, Hill’s constant of sigmoidicity, SITmi serum inhibitory titer term associated with 50%
enhancement of bacterial death, POPi is the estimated % of the ith population present in the initial inoculum
Discussion
• Drug effect gradually increases with increasing
concentration causing total inhibition of bacteria
replication
• The 2nd population was clearly seen by the inability of the
drug to have little to no effect on bacteria replication even
at its highest dose
• The poor effect on the 2nd population is reflected by the
high SITm2, and the presence of re-growth at peak
concentrations of 86.5XMIC for the X219 strain and
173.3XMIC for the X208 strain
Discussion
• Regrowth at high doses indicate the need for unsustainable
concentrations for substantial effect against the 2nd subpopulation
• The velocity of growth of the resistant population were 0.69X
and 0.71X that of the sensitive sub-population for the X208 and
X219 strains respectively indicating reduced fitness
• In- spite of β-lactamase activity, there is still a population that is
susceptible to cefprozil, is essentially ineffective against the 2nd
sub-population
Conclusion
• Cefprozil can be best modeled by inhibiting the replication of H.
flu
• The X204 strain can be best described by 1 sub-population,
and the X208 and X219 strains, by 2 sub-populations
• Cefprozil at adequate concentrations causes complete
inhibition of the bacteria replication
• Multiple regimens of ceprozil against H. flu could be well
described (comodeled) using a mathematical model
Questions ?