Comparing dosage adjustment methods for once-daily tobramycin in paediatric and adolescent patients with cystic fibrosis.

BACKGROUND AND OBJECTIVES: Several dosage adjustment methods are currently available to individualize intravenous tobramycin dosing. This study compared different methods in terms of their recommendations for dosage adjustment, their estimation of patients' pharmacokinetic parameter values and their ability to predict subsequent observed tobramycin concentrations following once-daily tobramycin treatment in children and adolescents with cystic fibrosis.
METHODS: Retrospective data from 172 patients treated at the Royal Children's Hospital (Brisbane, QLD, Australia) were analysed. To be included in the analysis, each patient had to have at least one pair of tobramycin plasma concentration-time measurements recorded over a dosing interval. One or both of the concentrations in the paired set were applied in each of the following dosage adjustment methods: (i) the Therapeutic Guidelines Aminoglycoside nomogram; (ii) the Massie nomogram; (iii) log-linear regression analysis; and two Bayesian forecasting software programs: (iv) TCIWorks and (v) DoseMe. All methods were compared in regard to their recommendations for tobramycin dosage adjustment. The latter three methods were also examined in terms of estimated pharmacokinetic parameter values and their ability to predict subsequent observed tobramycin concentrations.
RESULTS: The Therapeutic Guidelines nomogram recommended significantly greater mean doses for dosage adjustment (27.0 mg/kg) compared with all other methods (p ≤ 0.01), which gave similar mean dose recommendations (11.6-14.6 mg/kg); however, >20 % differences in doses on an individual level were seen on 20-35 % of occasions across all methods. The log-linear regression analysis method and the two Bayesian forecasting methods (TCIWorks and DoseMe) showed negligible bias but imprecision of around 20 % in predicting subsequent observed tobramycin concentrations. The Bayesian forecasting methods showed no significant difference in mean dose recommendations when using either one or two concentration measurements but increased imprecision in predicting subsequent observed tobramycin concentrations.
CONCLUSION: The log-linear regression method and Massie nomogram are likely to be suitable alternative methods for tobramycin dosage adjustment when Bayesian forecasting software is unavailable. The Therapeutic Guidelines nomogram should not be used to aid dose adjustment of tobramycin therapy in children with cystic fibrosis.