Fundamental problems with pediatric adaptive dosing of carboplatin using nuclear-medicine-based estimates of renal function

Background Carboplatin is often adaptively dosed based on glomerular filtration rate (GFR), usually estimated by nuclear medicine tests. At least five pediatric adaptive dosing formulas have been developed. In an effort to standardize dosing in Children's Oncology Group protocols, we explored methodologic variation in GFR estimation and adaptive-dosing formula performance. Procedure Nuclear medicine GFR data from published series of >= 100 children with cancer were compared. Data from patients for whom body surface area, weight, GFR, and tracer half-life were available were used to compare formulas. Results Differences in methods used to estimate GFR in children with cancer resulted in highly variable population results, with median GFRs ranging from 96 to 150 mL/min/1.73m(2). The choice of adaptive formula had a major impact on calculated dose. When targeting an area under the curve of 7.9 mg/mL center dot min, the median difference between the formula yielding the lowest and highest carboplatin dose for individual subjects was 289 (range 96-1 737) mg/m(2). Conclusions Wide variation in GFR obtained with nuclear-medicine-based tests in children with cancer primarily results from systematic methodologic errors. Formulas for calculating carboplatin dose produce additional and substantial variation that may place children with cancer at unnecessary risk for excessive toxicity or underdosing. These findings indicate a need for the development of a uniform, validated method for GFR determination in children that should be utilized in all centers. Currently, adaptive dosing of carboplatin based on GFR has serious limitations and in most clinical settings should arguably not be used in place of body-surface-area-based dosing.