Background and objective: The individual dosing of drugs that are mainly eliminated unchanged in the urine is made possible by assessing renal function. Most of the methods used are based on serum creatinine (SCr) levels. Cystatin C (CysC) has been proposed as an alternative endogenous marker of the glomerular filtration rate (GFR). Carboplatin is one of the drugs for which elimination is most dependent on the GFR. A prospective clinical trial including 45 patients was conducted to assess the value of serum CysC as a predictor of carboplatin clearance (CL).
Methods: The patients were receiving carboplatin as part of established protocols. Carboplatin was administered as a daily 60-minute infusion at doses ranging from 290 to 1700mg. A population pharmacokinetic analysis was performed using the nonlinear mixed effect modelling NONMEM program according to a two-compartment pharmacokinetic model.
Results: Data from 30 patients were used to test the relationships between carboplatin CL and morphological, biological and demographic covariates previously proposed for prediction of the GFR. The interindividual variability of carboplatin CL decreased from 31% (no covariate) to 14% by taking into account five covariates (SCr, CysC, bodyweight [BW], age and sex). Prospective evaluation of these relationships using the data from the other 15 patients confirmed that the best equation to predict carboplatin CL was based on these five covariates, with a mean absolute percentage error of 13% as an assessment of precision. NONMEM analysis of the whole dataset (n = 45 patients) was performed. The best covariate equation corresponding to the overall analysis was: CL (mL/min) = 110 x (SCr/75)-0.512 x (CysC/1.0)-0.327 x (BW/65)0.474 x (age/56)-0.387 x 0.854sex, with SCr in micromol/L, CysC in mg/L, BW in kilograms, age in years and sex = 0 if male and 1 if female. To put the value of CysC as an endogenous marker of the GFR into perspective, covariate equations without SCr were also evaluated; a better prediction was obtained by considering CysC together with age and BW (interindividual variability of 16.6% vs 23.3% for CysC alone).
Conclusion: CysC is a marker of drug elimination that is at least as good as SCr for predicting carboplatin CL. The model based on five covariates was superior to those based on only four covariates (with BW, age and sex combined with either SCr or CysC), indicating that CysC and SCr are not completely redundant to each other. Further pharmacokinetic evaluation is needed to determine whether SCr or CysC is the better marker of renal elimination of other drugs.