Disposition of Cefepime in the Central Nervous System of Patients With External Ventricular Drains

Denise H. Rhoney, Pharm.D., Vincent H. Tam, Pharm.D., Dennis Parker, Jr., Pharm.D., Peggy S. McKinnon, Pharm.D., William M. Coplin, M.D.


Pharmacotherapy. 2003;23(3) 

In This Article


Seven adults (five women, two men; median age 54 yrs; age range 41-79 yrs) with documented creatinine clearance greater than 60 ml/minute (median 71 ml/min, range 61-105 ml/min), calculated from a 24-hour urine collection, who received cefepime 2 g intravenously every 12 hours (each dose infused over 30 min) to treat nosocomial pneumonia were included in this open-label, prospective study. Patients were excluded for potentially altered pharmacokinetics (e.g., pregnancy, spinal cord injury), for weighing 40% over or under their ideal body weight, for fluctuating renal function (change in serum creatinine ± 20% of baseline), and for a current diagnosis of meningitis. Patients included in the study had external ventricular drains placed for medical reasons (i.e., increased intracranial pressure) before study enrollment. Most patients had an admission diagnosis of aneurysmal subarachnoid hemorrhage (five patients), followed by primary intracerebral hemorrhage (one) and penetrating head injury (one). The most common concomitant drugs received were tobramycin (seven patients), phenytoin (seven), nimodipine (six), and metoprolol (three). No adverse effects referable to cefepime were noted in any of the study patients. The Human Investigation Committee approved this study; informed consent was obtained from all subjects.

Serial serum and CSF samples were obtained concurrently after the fourth dose during one dosing interval. Three postinfusion concentrations were drawn as follows: 30 minutes after the infusion, at the dosing interval midpoint ± 1 hour, and 15-30 minutes before the next infusion. Samples were centrifuged and stored at -70°C.

Serum and CSF cefepime concentrations were determined by a validated high-performance liquid chromatography assay.[6] The lower limits of detection in serum and CSF were 0.5 and 0.25 µg/ml, respectively. The cefepime standard curves in serum and CSF were linear within the 0.5-50 and 0.25-30 µg/ml ranges, respectively. The interday coefficients of variation in serum of 2 µg/ml (10 samples) and 40 µg/ml (10 samples) were 3.4% and 4%, respectively, with an intraday and interday variability of less than 5%. The between-day coefficients of variation for CSF of 1 µg/ml (9 samples) and 25 µg/ml (9 samples) were 3.4% and 1.8%, respectively. The intraday variability of both serum and CSF was less than 4% across all concentrations.

The concentration-time profiles in serum and CSF were comodeled by using a two-compartment model with zero-order infusion to the central compartment (volume of distribution in the central compartment [Vc]) for each individual patient, with use of ADAPT II (Biomedical Simulations Resource, University of Southern California, Los Angeles, CA). Appendix 1 shows the differential equations for the proposed model. A separate variance model was used for the serum and CSF assay. Based on the model pharmacokinetic parameters derived, steady-state concentrations in serum and CSF were simulated. Each pharmacokinetic parameter was analyzed and reported by the standard two-stage method. Areas under the concentration-time curves (AUCs) in CSF and serum were derived by integrating the respective concentration-time profiles with respect to time during one dosing interval (12 hrs). The volume of the peripheral compartment was assumed to be constant, and mass transfer or distribution within the compartment was assumed to be instantaneous. We did not assume that equilibrium had been achieved in the peripheral compartment in our pharmacokinetic analysis. For each patient, the time course of the doses administered and the sampling times in relation to the last dose given were taken into account to derive the model parameters. The steady-state drug exposures reported were simulated with numerous (> 10) doses, and the achievement of steady state was confirmed by comparing various concentrations or exposures with consecutive doses in the simulations. The CSF concentration-time profiles of the individual patients were compared with published MIC90 of common pathogens isolated in patients with nosocomial meningitis.[7]