Our previous study indicates that cerebrospinal fluid (CSF) albumin level is a determinant of CSF vancomycin concentration for postoperative neurosurgical patients. We aimed to develop an improved vancomycin population pharmacokinetic model with incorporation of more covariates, and to provide dosing guidance for clinicians. Vancomycin was administered intravenously to 20 patients with external ventricular drains after neurosurgical operation. Blood and CSF were collected and vancomycin concentrations were measured by HPLC. A separate CSF compartment was considered, and was linked to the central compartment by a first-order process (Q(CSF)). The clearance of the CSF compartment (Cl-CSF) was used to characterize vancomycin elimination from CSF through external ventricular drain. Nonlinear mixed-effects modeling approach was used to develop the model. The CSF albumin level (mg/dL) was the covariate influencing Q(CSF): Q(CSF) = 0: 0049 + 0: 000021 x (CSF albumin - 279). The effect of body weight (BW, kg) was significant on central volume (V-C): V-C = 27: 84 + 0: 96 x (BW - 69). All parameters were estimated with an acceptable precision (relative standard error: RSE% < 30.26). The performance of the final model was acceptable with our previous dataset. A simple to use dosage regimen table was created to guide clinicians with vancomycin dosing. This model incorporates variables of both CSF albumin and BW, which offers improvements to the previous pharmacokinetics model. (C) 2016 Published by Elsevier Inc. on behalf of the American Pharmacists Association.