Use of gentamicin serum levels to individualize therapy in children

Abstract

The value of a one-compartment pharmacokinetic model for calculating gentamicin half-life and predicting steady-state serum concentrations was assessed in 20 children, 32 to 168 months of age, who were receiving the drug intravenously for gram-negative bacterial infections. Actual steady-state serum concentrations measured after the twenty-fourth or twenty-fifth doses were compared with corresponding values predicted by the model. The mean predicted steady-state serum concentration at 30 minutes after the intravenous dose, 3.6 μg/ml, was 13% less than the mean measured value, 4.15 μg/ml. The mean predicted steady-state nadir serum concentration was 0.152 μg/ml, or 54% less than the mean measured value, 0.33 μg/ml. The mean t 1/2 for decline in serum concentrations following the first dose was 66.9 minutes, which was significantly different ( P 1/2 at steady state (84.9 minutes). Serum concentrations measured in four patients during the washout phase following the last dose of gentamicin declined in a biphasic manner, indicating two-compartment distribution of the drug. The error in predicted-versus-measured serum levels and the change in t 1/2 resulted from the use of the currently accepted but inappropriate one-compartment model to characterize gentamicin disposition. The change in t 1/2 with multiple dosing was attributable to tissue accumulation and not to impaired excretion of gentamicin. Alternative methods to reduce the error in characterizing gentamicin pharmacokinetics and to improve the utility of measuring gentamicin serum levels are proposed.