Vancomycin dosing in children and young adults: back to the drawing board.

STUDY OBJECTIVE

To compare the likelihood of alternative vancomycin dosing strategies based on weight, height, or body surface area (BSA) in achieving isometric vancomycin area under the serum concentration-time curve (AUC) values across the body size distribution of children and young adults.

DESIGN

Maximum a posteriori probability Bayesian (MAP-Bayesian) pharmacokinetic analysis using retrospectively collected medical record data.

SETTING

Children's hospital.

PATIENTS

A total of 115 patients 1-20 years of age managed outside of the intensive care unit who were treated with vancomycin between May 1, 2011, and August 31, 2012, and had a minimum of two serum vancomycin concentration measurements.

MEASUREMENTS AND MAIN RESULTS

Vancomycin dosing and concentration-time information along with demographic and laboratory data related to kidney function estimation were extracted from the patients' medical records. A previously structured one-compartment model was used to derive MAP-Bayesian estimates of pharmacokinetic system parameters for each patient. Post hoc linear and power function regression were used to compare clearance (Cl) and the volume of distribution of the central compartment (Vc) to body size descriptors (weight, height, and BSA). The relationship of the body size descriptor-indexed parameter across the body size distribution was assessed. The AUC from time 0 to 24 hours (AUC24 ) values associated with vancomycin dosing regimens based on weight, height, and BSA were estimated. The 115 patients (56.5% male) had a mean ± SD age, height, and weight of 9.7 ± 5.4 years, 133 ± 32.3 cm, and 38.0 ± 24.2 kg, respectively. Each patient received a minimum of four doses of vancomycin and had two to nine serum vancomycin concentration measurements, for a total of 313 measurements for all patients. Vancomycin Cl was a nonlinear function of weight and a linear-proportionate function of BSA, whereas the volume of distribution of the central compartment (Vc) was a linear-proportionate function of weight. The expected median (5th-95th percentile) AUC24 values with weight-based dosing of vancomycin 60-70 mg/kg/day were 446 (315-834) mg·hour/L and 649 (385-1165) mg·hour/L in patients weighing less than 40 kg (n=72) and those weighing 40 kg or more (n=43), respectively. In contrast, isometric AUC24 values were predicted with BSA-based dosing across the body size distribution.

CONCLUSION

BSA-based dosing is more likely than weight-based (mg/kg) dosing of vancomycin to achieve isometric AUC24 values across the body size distribution of children and young adults. Pharmacokinetic studies that compare these two vancomycin dosing strategies in children are clearly needed to validate these findings.