Antibacterial therapeutic drug monitoring in cerebrospinal fluid: difficulty in achieving adequate drug concentrations.

Dagan O. Lonsdale, Andrew A. Udy, Jason A. Roberts, Jeffrey Lipman

This report illustrates the difficulty in managing CNS infection in neurosurgical patients, the altered drug pharmacokinetics associated with critical illness, and the role that therapeutic drug monitoring (TDM) of CSF can play in assisting clinical decision making. The authors present a case of external ventricular drain-related ventriculitis in a critically ill patient who initially presented with a subarachnoid hemorrhage. They discuss the physiological changes found in such patients, in particular augmented renal clearance (demonstrated in this patient by a measured creatinine clearance of 375 ml/min/1.73 m(2)), noting the effect this had on drug pharmacokinetics and leading to dosing requirements 2-3 times those recommended in standard regimens. The authors consider the bacterial "kill" characteristics of 2 different antibacterial agents (meropenem and vancomycin) and describe the unique approach of using plasma and CSF TDM to achieve optimal drug exposure at the site of infection while limiting toxic side effects. The authors demonstrate that simply using plasma TDM as a surrogate marker for drug concentration in the CNS may lead to underdosing, exemplified in this patient by CSF vancomycin concentrations as little as 13% of that in plasma. Finally, by measuring CSF and plasma ratios, the authors illustrate the disparity in pharmacokinetic properties between drugs, reminding the clinician of the importance of CNS penetration when selecting antibacterial agents in such cases. This work raises an important hypothesis in the accurate prescription of antibacterial agents in neurosurgical critical care, namely underdosing in the context of augmented elimination and impaired target site penetration. However, prior to any recommendations regarding empirical dose modification, more data are clearly needed, particularly with respect to the safety and efficacy of such an approach. In this respect, the authors would advocate further research using TDM in the management of CNS infection in this setting, in addition to work defining plasma and CSF concentrations associated with antibacterial efficacy and toxicity.