Background: Although aminoglycosides are routinely used in neonates, controversy exists regarding empiric dosing regimens. The objectives were to determine gentamicin pharmacokinetics in neonates, and develop initial mg/kg dosing recommendations that optimized target peak and trough concentration attainment for conventional and extended-interval dosing (EID) regimens.
Methods: Patient demographics and steady-state gentamicin concentration data were retrospectively collected for 60 neonates with no renal impairment admitted to a level III neonatal intensive care unit. Mean pharmacokinetics were calculated and multiple linear regression was performed to determine significant covariates of clearance (L/h) and volume of distribution (L). Classification and regression tree (CART) analysis identified breakpoints for significant covariates. Monte Carlo Simulation (MCS) was used to determine optimal dosing recommendations for each CART-identified sub-group.
Results: Gentamicin clearance and volume of distribution were significantly associated with weight at gentamicin initiation. CART-identified breakpoints for weight at gentamicin initiation were: ≤ 850 g, 851-1200 g, and > 1200 g. MCS identified that a conventional dose of gentamicin 3.5 mg/kg given every 48 h or an EID of 8-9 mg/kg administered every 72 h in neonates weighing ≤ 850 g, and every 24 and 48 h, respectively, in neonates weighing 851-1200 g, provided the best probability of attaining conventional (peak: 5-10 mg/L and trough: ≤ 2 mg/L) and EID targets (peak:12-20 mg/L, trough:≤ 0.5 mg/L). Insufficient sample size in the > 1200 g neonatal group precluded further investigation of this weight category.
Conclusions: This study provides initial gentamicin dosing recommendations that optimize target attainment for conventional and EID regimens in neonates weighing ≤ 1200 g. Prospective validation and empiric dose optimization for neonates > 1200 g is needed.
Keywords: Gentamicin; Neonate; Pharmacokinetics; Traditional dosing, extended-interval dosing, Monte Carlo simulation.