Aims: Physiologic changes due to disease status can lead to highly variable pharmacokinetics (PK) of amikacin in children. Therefore, considerations should be given when determining optimal amikacin dose for patients with specific diseases. Improved understanding of factors influencing PK can allow for the optimizing of dosage regimens to reduce adverse effects. The aim of the study was to undertake a comparative pharmacometric analysis of amikacin use in pediatric patients with burn injuries verses those with oncology conditions.
Methods: Population PK modeling was performed in NONMEM® version 7.3 (ICON Development Solutions, Ellicott City, MD, USA) for pediatric patients < 18 years of age with burn injuries (group 1) and with oncology conditions (group 2) who received > 1 dose of amikacin for the empirical treatment of gram-negative bacterial infections between Jan 2007 to Dec 2009 and Jan 2006 to Dec 2011, respectively. The covariates examined included sex, body weight (WT), age (AGE), height, serum creatinine (SCR), C-reactive protein, and serum albumin levels. Both studies used the enzymatic method to measure SCR. The Schwartz equation was used to calculate creatinine clearance for both groups. The available shared covariates between the two subpopulations were separately checked in individual models. A binary covariate TYPE was created to denote the two different patient groups (burn 0, oncology 1). Patients with burn injuries received amikacin at 10-20 mg/kg/day as part of early empiric treatment of presumed burn-related sepsis. Pediatric oncology patients received 20 mg/kg of amikacin intravenously for treatment of febrile neutropenia. Patients were treated once daily and concentrations were collected immediately before and 1 hour after the second dose. A population PK model was developed to describe the observed amikacin concentration over time in both subpopulations.
Results: Data from pediatric patients (0.83 – 15 years) with burn injuries (n=72) and pediatric patients (0.8 – 16.4 years) with oncology conditions (n=111) were pooled for model development. The characteristics between the two groups was very similar with no statistical differences. In the burn injury group and the oncology group, SCR had a median (range) of 0.4 (0.2 – 1) and 0.32 (0.15 – 0.68) mg/dL, respectively. In addition the median of creatinine clearance (CRCL) was 170.5 and 195.4 mL/min/1.73m2, respectively. However, the overall range was lower in the burn injury patient group, 81.1 – 327.5 mL/min/1.73m2. The data were best described by a two-compartmental PK model with a proportional error model. Stepwise covariate search (forward addition p<0.05, backward elimination p<0.01) returned WT, subpopulation type (TYPE), AGE and CRCL as significant covariates. Allometric scaling was initially tested during model development but was abandoned after systematic covariate analysis. The final population PK covariate model is outlined in Table 1.
Table. 1 Summary of PK parameters
Conclusion: The results of the current study suggest that besides patient-specific characteristics (current WT, AGE and CRCL) also disease-related characteristics should be considered when dosing amikacin in critically ill pediatric patients, in order to optimize therapeutic concentration targeting.