A retrospective analysis using Monte Carlo simulation to evaluate recommended ceftazidime dosing regimens in healthy volunteers, patients with cystic fibrosis, and patients in the intensive care unit.

BACKGROUND

Over the past decades, the relationship between the pharmacokinetic (PK) properties of antibiotics, MICs, and clinical effects has been increasingly well understood. Interpatient variability in the PK profile, however, has only recently been recognized as a major factor in predicting the outcome in individual patients and establishing breakpoints for clinical susceptibility. Most predictions to date have used data from healthy volunteers.

OBJECTIVE

The purpose of this study was to perform Monte Carlo simulations of the PK/pharmacodynamic relationships of ceftazidime to assess whether the probability of target attainment (PTA) differed significantly between 3 distinct populations. To that end, population PK models of ceftazidime were developed for the 3 populations.

METHODS

Serum concentration-time data from earlier studies in healthy volunteers (n = 8), patients with cystic fibrosis (CF) (n = 17), and patients in the intensive care unit (ICU) (n = 6) were used to obtain population PK parameter estimates and covariance matrices using the nonparametric adaptive grid program. The PTA for each group was obtained using 10,000 patient simulations for dosing regimens of 1000 and 2000 mg q8h over a range of MICs and percentages of time that concentrations of unbound drug remained above the MIC (%T > MIC).

RESULTS

The relationship between the MIC and the population mean %T > MIC, as well as the PTA profiles, differed markedly between the 3 groups as a result of both differences and variations in V(d) and Cl. Breakpoints based on a 100% PTA for a %T > MIC of 60% were < or = 4, 0.5, and 0.5 mg/L in healthy volunteers, patients with CF, and patients in the ICU, respectively. However, when PTA values between 90% and 100% were reevaluated and differences in clinical dosing regimens were accounted for, the resulting breakpoints were identical in the 3 groups.

CONCLUSIONS

PK parameter estimates for ceftazidime based on data from a small group of healthy volunteers resulted in a clinical susceptibility breakpoint comparable to those for patients with CF and patients in the ICU. Based on the study findings, this breakpoint would be < or = 4 mg/L. Patients suspected of having unusually high rates of clearance should be monitored closely.