The purpose of this study was to describe the maturation of vancomycin (V) clearance and the influence of altered renal function in infants on vancomycin using population pharmacokinetic methods. A population pharmacokinetic model was developed using NONMEM from clinical data obtained from 374 newborns and infants < 2 years of age (median age = 27 days) from four institutions. A total of 1103 serum V concentrations were used in the model development, including 311 with elevated serum creatinine (CR) (> 0.8 mg/dl) and more than 104 evaluations in infants older than 2 months of age. The final model was evaluated against a second data set of 160 concentrations from 67 infants at one of the institutions and then used to develop dosing guidelines. The data were best described by a two-compartment model. Weight and CR greatly influenced vancomycin elimination, while postnatal age and prematurity (< 28 weeks) were significant but less important predictors of V elimination. For the typical study infant (age = 27 days, CR = 0.6, WT= 1.8 kg, gestational age = 33.5 weeks), this results in VdSS = 0.79 l/kg and Cl = 0.066 l/h/kg. The validation data set showed the model to be unbiased. Dosing guidelines from this model, based on serum creatinine and gestational age at birth, performed better than published guidelines based on postconceptional age. Vancomycin clearance is initially reduced in premature infants and increases with postnatal age. Most of the age-related changes could be predicted by the concomitant fall in serum creatinine. Dosing guidelines that incorporate these factors are more likely to produce therapeutic V concentrations in infants.