Vancomycin pharmacokinetics in neonates and infants: a retrospective evaluation

To evaluate the frequency with which current loading and maintenance vancomycin dosages achieve target serum concentrations based on pharmacokinetic parameters obtained after the initial dose. Also, to identify the daily vancomycin dosage necessary to achieve target serum concentrations at steady-state and to determine if any relationships exist between vancomycin pharmacokinetic parameters and various patient characteristics.

Neonatal intensive care unit (NICU) at Georgia Baptist Medical Center.

Twenty-three infants with suspected or documented gram-positive infection who received intravenous vancomycin between July 1990 and November 1991 were included in this retrospective analysis. Gestational age range from 23 to 41 weeks and postconceptional age (PCA) at the time of the study ranged from 26 to 46 weeks. Vancomycin therapy was initiated with a loading dose of 15 mg/kg, followed by a maintenance dosage of 20-30 mg/kg/d, which was usually given as 10 mg/kg q8-12h. All vancomycin doses were administered using a syringe pump. Peak and trough serum concentrations were obtained following the first dose. Vancomycin pharmacokinetic parameters were determined using a one-compartment model. Infants receiving indomethacin within two weeks prior to study were analyzed separately (group 2, n = 4). All other infants were included in group 1 (n = 19).

For group 1, vancomycin clearance (Cl), volume of distribution (Vd), and half-life were (mean +/- 1 SD) 0.072 +/- 0.032 L/kg/h, 0.52 +/- 0.08 L/kg, and 5.6 +/- 1.6 hours, respectively. For both groups, loading doses provided 1-hour postinfusion peak concentrations of 25-35 mg/L in one of every two infants studied, whereas only three percent of initial maintenance doses were projected to provide desired peak and trough concentrations at steady-state. For group 1, the mean daily dosage necessary to provide target peak (25-35 mg/L) and trough (5-10 mg/L) concentrations at steady-state was larger than that initially prescribed (29.6 +/- 13.1 vs. 22.2 +/- 4.7 mg/kg/d). For group 2, the mean daily dosage required to achieve target peak and trough concentrations at steady-state was smaller than that initially prescribed (14.8 +/- 4.3 vs. 20.0 +/- 0.1 mg/kg/d) and was exactly half of that required for group 1. Excellent correlations were observed between PCA and vancomycin Cl (L/h) (r = 0.92; p < 0.0001), body weight and Vd(L) (r = 0.94; p < 0.0001), body weight and vancomycin Cl (L/h) (r = 0.85; p < 0.0001), PCA and Vd (L) (r = 0.89; p < 0.0001), and body surface area and Vd (L) (r = 0.93; p < 0.0001) for group 1. Moderate correlations were also noted between PCA and Cl relative to body weight (L/kg/h), postnatal age and Cl (L/kg/h), and PCA and vancomycin dosage requirements (mg/kg/d). No linear correlation was observed between any patient characteristic and Vd standardized for body weight.

Our data demonstrate the need for a more accurate method of estimating initial vancomycin dosage requirements in this NICU population. Although some of the relationships revealed in this study could be used to determine vancomycin dosage for infants in the range of approximately 30-36 weeks PCA, we hesitate to suggest this approach presently because of the potential limitations of our study design. Further prospective study is needed to confirm these observations. In addition, further study is necessary to describe the time course of the interaction between vancomycin and indomethacin in infants with successful and unsuccessful closure of their patent ductus arteriosus.