Icaac04- a-138

ICAAC 2004
Use of Pharmacokinetic-Pharmacodynamic and Monte Carlo Simulation as Decision Support
for the Re-Evaluation of NCCLS Cephem Susceptibility Breakpoints for Enterobacteriaceae
PG AMBROSE, SM BHAVNANI, RN JONES, WA CRAIG, MN DUDLEY
Cognigen Corp., Buffalo, NY; University at Buffalo, Buffalo, NY; JONES Group/JMI Laboratories, North Liberty, IA;
University of Wisconsin, Madison, WI; and Diversa Corporation, San Diego, CA
Animal PK-PD Models
Current, PK-PD-based, PK-PD-based-population distribution adjusted breakpoints A M E N D E D A B S T R A C T
¥ Neutropenic murine-thigh infection models were used to study PK-PD indices that correlated for antimicrobial agents in NCCLS M100-S14 Table 2A.
best with efficacy for Enterobacteriaceae, including both EBSL- and non-ESBL producing Background: As cephem-resistant Enterobacteriaceae, including ESBL-producing strains,
have continued to emerge, the appropriateness of current NCCLS susceptibility breakpoints Potential susceptibility breakpoints could be stratified into one of five groups Table 2 shows current cephem NCCLS susceptibility breakpoints, PK-PD based have been the focus of recent attention. In the face of limited clinical data, Pharmacokinetic- ¥ Indices of efficacy explored include: Duration of time (T) drug concentration remain above the based upon probability of PK-PD target attainment and examination of breakpoints, and MIC population-adjusted PK-PD breakpoints.
Pharmacodynamics (PK-PD) and Monte Carlo simulation have been increasingly used as MIC (T > MIC), ratio of the peak concentration of the agent to the MIC of the pathogen (Cmax:MIC decision support for establishing breakpoints. NCCLS susceptibility breakpoints for cephems ratio), ratio of the 24-hour area under the concentration-time curve (AUC0-24) of the agent to against Enterobacteriaceae were re-evaluated using this approach.
Figure 3. Fractional PK-PD target attainment of cefuroxime against Escherichia coli and
Methods:
Susceptibility breakpoints did not bisect MIC distribution (Example, Figures Using published mean PK parameter estimates and dispersion measures from . The probability of PK-PD target attainment is greater than ¥ Non-linear regression using a Hill-type model was used to describe relationship between the volunteer studies and PK-PD targets derived from a murine infection model (% free-drug (f) 2A and 2B): Ceftazidime, Ceftriaxone, Cefotaxime, Cefepime, Cefamandole, 0.9 up to MIC values of 4 mg/L when cefuroxime is dosed 0.75 g every 8 hours above PK-PD measures and response to therapy.
time > MIC ~50), Monte Carlo simulation (10,000 subjects) was used to evaluate the probability and 8 mg/L when dosed 1.5 g every 8 hours.
of PK-PD target attainment (³ 90%) for standard cephem dosing regimens.
¥ Figure 1 shows the relationship between cephem exposure, as measured by %T > MIC, and Results: Simulation results support lowering susceptibility breakpoints 2-16-fold for most
response in neutropenic murine-thigh infection models involving Gram-negative bacteria.
Susceptibility breakpoints bisecting MIC population distributions could be agents (see table below). Except for cefoxitin and cefazolin, breakpoints do not bisect - 50% T>MIC was associated with a 0.8 to 2.4 log-unit reduction in bacterial burden.
avoided assuming higher of labeled doses used for serious infections contemporary MIC population distributions (SENTRY Program).
Monte Carlo Simulation
¥ The following PK-PD structural model was used in the simulations: T >MIC = (ln Dose/(V§/f) Ð ln MIC)/(CLT/V§) where V§ is the volume estimated from the MIC Probability
Susceptibility breakpoints bisect MIC distribution (Example, Figure 4): terminal elimination phase half-life (§) for a two-compartment model, CL f is the fraction of unbound drug.
Probability of Target Attainment
¥ Pharmacokinetic parameters for each simulated patient were based on mean pharmacokinetic Contemporary MIC populations not available: Ceftizoxime, Cefotetan, MIC (mg/L)
parameter estimates and measures of dispersion.
- A log-normal distribution was assumed for clearance and volume terms.
¥ 10,000 patients were simulated for each drug regimen.
Consider removal from NCCLS table 2A due to lack of use and unavailability Figure 4. Fractional PK-PD target attainment of cefoxitin against Escherichia coli and Klebsiella
- Simulations were conducted using the computer software Crystal Ball 2000.1 by pneumoniae. The probability of PK-PD target attainment is greater than 0.9 at a of contemporary MIC data: Cephalothin, Cefonicid, Moxalactam C O N C L U S I O N S
MIC of 2 mg/L when cefoxitin is dosed 1 g every 6 hours and 4 mg/L when dosed Susceptibility Breakpoint Determinations
Figure 2A. Fractional PK-PD target attainment of ceftriaxone against Escherichia coli and
¥ 50% T>MIC was utilized as the PK-PD goal of therapy.
Klebsiella pneumoniae. The probability of PK-PD target attainment is essentially Simulation results suggest that susceptibility breakpoints for 1.0 up to MIC values of 1 mg/L when ceftriaxone is dosed 1 g every 24 hours and Conclusions: Simulation results suggest that susceptibility breakpoints for most cephems
¥ The susceptibility breakpoint was the highest MIC value associated with at least a 90% most cephems against Enterobacteriaceae should be reduced 2 mg/L when dosed 2 g every 24 hours.
against Enterobacteriaceae should be reduced to insure that most patients attain PK-PD probability of PK-PD target attainment.
to insure that most patients attain PK-PD exposures associated ¥ Contemporary MIC population distributions were considered (Escherichia. coli and Klebsiella MIC Probability
I N T R O D U C T I O N
- If a major portion of the MIC distribution was bisected by a PK-PD susceptibility breakpoint, Probability of Target Attainment
alternative breakpoints were chosen to avoid division but ensure high probabilities of PK- As cephem-resistant Enterobacteriaceae, including ESBL-producing strains, have continued to PD target attainment. In these cases, only susceptible and resistant breakpoints chosen.
Adjustment of NCCLS breakpoints to lower values may negate MIC (mg/L)
MIC Probability
emerge, the appropriateness of current NCCLS susceptibility breakpoints have been the focus the clinical need for ESBL screening and confirmatory tests.
Probability of Target Attainment
Figure 1.
Relationship between T>MIC and change in bacterial density in thighs of mice at - Re-evaluation of susceptibility breakpoints for MIC and Information from animal infection models, as well as limited clinical data, suggest that ESBL- 24 hours. Each data point represents data for one mouse. The dotted line reflects Pharmacokinetic parameters utilized for simulations.
MIC (mg/L)
producing strains with lower MIC values may be effectively treated using appropriate cephem the number of bacteria at the beginning of therapy. The T>MIC required for efficacy corresponding disk zone diameters to better predict is not affected by ESBL-production status.
Pharmacokinetic Parameter Estimates (mean (SD or range)) presence of ESBLs while potentially eliminating separate Figure 2B. Fractional PK-PD target attainment of cefepime against Escherichia coli and
Given the paucity of clinical data, pharmacokinetics-pharmacodynamics (PK-PD) and Monte Klebsiella pneumoniae. The probability of PK-PD target attainment is greater than screening tests has been completed (see Poster D-303, Carlo simulation have been increasingly used as decision support for establishing breakpoints.
0.9 for MIC values of 4 mg/L irrespective of the three regimens considered. When cefepime is dosed 2 g every 12 hours or 1 g every eight hours, the probability of In view of the above limitations, susceptibility breakpoints for cephems against Enterobacteriaceae PK-PD target attainment is 0.9 and 0.77 for a MIC of 8 mg/L, respectively.
were re-evaluated using this approach.
- Current NCCLS Enterobacteriaceae susceptibility breakpoints remain acceptable based on error rate analysis M AT E R I A L S A N D M E T H O D S
for contemporary isolates. Lower breakpoints could eliminate Pharmacokinetic Parameter Estimates
¥ Pharmacokinetic data from healthy volunteers were extracted from the medical literature the routine need for separate ESBL screening tests and MIC Probability 0.3
disk diffusion zone correlates can be selected to maximize Probability of Target Attainment
inter-method accuracy pending final approval by the NCCLS.
Microbiological Susceptibility Data
MIC (mg/L)
¥ Obtained from the SENTRY Antimicrobial Surveillance Program (2001-2003).

Source: http://www.icpd.com/files/ICAAC_04_A_138_Enterobacteriaceae_ceph_breakpoints.pdf